Package Details: pi-hole-server 5.11.4-3

Git Clone URL: (read-only, click to copy)
Package Base: pi-hole-server
Description: The Pi-hole is an advertising-aware DNS/Web server. Arch adaptation for lan wide DNS server.
Upstream URL:
Keywords: ad block pi-hole
Licenses: EUPL-1.2
Conflicts: pi-hole-standalone
Submitter: max.bra
Maintainer: max.bra (graysky)
Last Packager: max.bra
Votes: 97
Popularity: 0.85
First Submitted: 2016-01-13 12:50 (UTC)
Last Updated: 2022-08-07 19:01 (UTC)

Pinned Comments

max.bra commented on 2018-02-09 16:45 (UTC) (edited on 2019-10-18 23:14 (UTC) by max.bra)

ArchLinux Pi-hole is not officially supported by Pi-hole project. In case of bugs and malfunctions please DO NOT file a report upstream.

First of all check if the wiki ( can help then ask here for assistance and tips.
When it will be excluded that the problem does not depend on ArchLinux we will file a bug upstream.

Latest Comments

max.bra commented on 2022-08-07 19:03 (UTC)

@Mettacrawer thanks for report and for the upstream patch. early applied.

Mettacrawer commented on 2022-08-07 09:20 (UTC)

This does not concern this AUR package. If you are getting a lot of log messages that go "FastCGI-stderr:PHP Notice: Undefined index: client_id in /srv/http/pihole/admin/api_db.php on line 127" there was an upstream bug that with a fix that had not been released yet. I patched api_db.php on my Pi-hole the same way as in the PR and it worked.

max.bra commented on 2022-07-13 20:06 (UTC) (edited on 2022-07-13 20:31 (UTC) by max.bra)

@adityaruplaha problem applying what setting?

Edit: nevermind. solved. thanks for reporting.

adityaruplaha commented on 2022-07-13 18:22 (UTC)

Web settings are broken as of now. The /run files are in a different location.

There was a problem applying your settings.
Debugging information:
PHP error (2): file_get_contents(/run/pihole-FTL.port): Failed to open stream: No such file or directory in /srv/http/pihole/admin/scripts/pi-hole/php/FTL.php:36

However, changing the location does not fix things. Settings refuse to be applied.

max.bra commented on 2022-07-05 14:18 (UTC)

hi Adamsdesk. Patched: you should safely remove your symlink. Thanks a lot for reporting.

Adamsdesk commented on 2022-07-04 23:23 (UTC) (edited on 2022-07-04 23:27 (UTC) by Adamsdesk)

I recently setup a new install of pi-hole-server 5.10-1 using Nginx and noticed that the teleporter is not working from the command line (pihole -a -t) due to an invalid path being referenced. From what I've found the pihole command calls upon the /opt/pihole/ which has an invalid path hard coded as follows.

653:    php /var/www/html/admin/scripts/pi-hole/php/teleporter.php > "${filename}"

My quick work around was to create a directory and then a symbolic link as follows.

# mkdir /var/www
# ln -s /srv/http/pihole/ /var/www/html

Note: Using teleporter via the web interface works fine.

max.bra commented on 2022-05-14 12:42 (UTC)

hi kuba-orlik, FTL updated. Thanks for reporting.

kuba-orlik commented on 2022-05-14 06:28 (UTC)

I've encountered an issue with DHCP:

Running install -m 644 -o pihole -g pihole /dev/null /etc/pihole/dhcp.leases fixed the UI issue. Not sure why this file was missing.

I got the inspiration to fix it that way by studying this file:

max.bra commented on 2022-03-24 17:12 (UTC)

hi dmellado, please can you read posts from m_st (just before yours) and have your say?

dmellado commented on 2022-03-23 10:38 (UTC) (edited on 2022-03-23 10:39 (UTC) by dmellado)

Hi! I've been checking that using the latest package compiled I'm getting duplicated entries in 01-pihole.conf from the upstream DNS'. Rather than removing them it's just stacking them out in the file.


# If a DHCP client claims that its name is "wpad", ignore that.
# This fixes a security hole. see CERT Vulnerability VU#598349

max.bra commented on 2022-03-14 09:05 (UTC)

@m_st i don't think so, but anyway try: from the snapshot use dnsmasq.include as your "new default" 01-pihole.conf replacing yours and see what happens.

m_st commented on 2022-03-14 08:59 (UTC)

@max.bra No, I didn't find out. Maybe it was caused by editing the file manually before?

max.bra commented on 2022-03-06 16:57 (UTC)

@m_st mmh, no duplication here... can you find the triggering cause? I mean if is there a setting (e.g. Allow only local requests) which if checked triggers duplication...

m_st commented on 2022-03-06 16:52 (UTC)

@max.bra The new package did write to the config file, while still duplicating the entries in there. But now everything seems to be working and it doesn't use the google DNS anymore. Thanks for your help!

max.bra commented on 2022-03-06 11:39 (UTC)

@jxir no I was not!! Thanks for reporting. I think i've issued a CTRL+D instead of CTRL+S...

jxir commented on 2022-03-06 11:33 (UTC)

Hi @max.bra, I think you didn't mean to remove the depends variable in revision 3 ?

max.bra commented on 2022-03-06 10:42 (UTC)

yes it's a sudoers related permission problem. please try revision 3.

m_st commented on 2022-03-06 10:18 (UTC)

@max.bra That's what I thought aswell, I didn't see this file before and only used the gui. The file looks exactly like that.

I configured it in the web interface and restarted pi-hole, but this didn't change 01-pihole.conf. Also I now got this error: Warning in dnsmasq core: no upstream servers configured

So it seems the web interface does not (or cannot) write anything to this config file. However, the selection of upstream servers in the web interface is still being saved (I guess somewhere in the database?).

Should pi-hole actually write to this config file? Then maybe there is a permission problem?

max.bra commented on 2022-03-06 10:01 (UTC)

01-pihole.conf seems to be a bit messed up. Was it a problem in the copy paste or do I really see a lot of duplicate entries? ... and there is also a google server (, the one you were probably looking for. Try to delete everything underneath:

# If a DHCP client claims that its name is "wpad", ignore that.
# This fixes a security hole. see CERT Vulnerability VU#598349

and reconfigure DNS via web interface.

m_st commented on 2022-03-06 09:51 (UTC) (edited on 2022-03-06 09:51 (UTC) by m_st)

@max.bra Thanks for your quick response! I'm using the default installation which is using NetworkManager. The NetworkManager.conf file is empty. The connection file looks like this:

This is my resolv.conf:

Here is my 01-pihole.conf:

I hope that helps. Please tell me if you need anything else.

max.bra commented on 2022-03-05 12:41 (UTC) (edited on 2022-03-06 10:02 (UTC) by max.bra)

Hi m_st, you should provide everything about your network manager / configuration right now (resolv.conf, network manager conf files, systemd related...)

m_st commented on 2022-03-05 12:27 (UTC)

Hi @max.bra, I recently installed pi-hole on my private server running Manjaro linux. I set it up to only use cloudflare servers as upstream dns in the web interface. However, for the majority of requests, pi-hole uses the google dns to respond to requests. I filed a git issue here with some more information (knowing now that I shouldn't have filed it):

I didn't find any help in the comments, on github or the arch wiki page. Which additional information do you need?

max.bra commented on 2022-02-16 21:26 (UTC)

hi jxir. added, thanks for reporting.

jxir commented on 2022-02-16 21:11 (UTC)

> pihole enable
/usr/bin/pihole: line 25: /opt/pihole/ No such file or directory

The file is missing from the package. Adding the corresponding install command fixed the problem for me. Possibly more files are missing.

max.bra commented on 2022-01-06 11:09 (UTC)

@XenHat yes nice and good point but we cannot patch upstream source code to remove a core dependency... this goes a little further an AUR package, I think... don't you?

xenhat commented on 2022-01-05 21:54 (UTC)

In response to the core/inetutils dependency pointed out by @kusky, it can be avoided by patching to use hostnamectl hostname instead which returns the same value.

max.bra commented on 2022-01-04 13:40 (UTC) (edited on 2022-01-04 18:25 (UTC) by max.bra)

@starquake if you take a look at various patches applied to ftl, core and web admin you will see that we try to optimize the experience (like SD write cicle) adapting pi-hole in an arch way. Using the bin version will reset all our work.

df8oe commented on 2022-01-04 13:19 (UTC) (edited on 2022-01-04 13:20 (UTC) by df8oe)

I am working also with RPI 3B+. I do not have any problems with pihole performance - even if it is compiling. Process of compiling takes some minutes - but you cannot see any impact on running pihole. Maybe there are other processes (like a GUI) eating ressorces. I have located the out-of-memory-issue when compile pi-hole-server by watching 'top' on another console. You can see there memory is going lower and lower and if it reaches zero compiler shuts down. If you have performance problems with pi-hole it would be a good step to look at normal operation conditions using 'top'.

starquake commented on 2022-01-04 13:09 (UTC)

@max.bra I understand building from sources is recommended. And I agree with that. But some AUR packages have a -bin package for convenience. So that's why I asked. Because I would find that convenient. But maybe that's not a concern for most people.

Swapping will help but the machine gets really slow sometimes. And then pihole gets slow. I'll see if I can do cross compiling. Or maybe upgrade.


max.bra commented on 2022-01-04 08:46 (UTC)

@starquake hi, FTL has compile problems on systems with less than about 800MB of free RAM.
The binaries are already made available by the original developers but archlinux policies, where possible, recommend compiling the sources.
You can brilliantly solve the problem by configuring even only 512MB of swap or if possible do a cross compiling with your other device on the network (plus: also speeding things up).

df8oe commented on 2022-01-04 08:41 (UTC)

I do not think a binary package is necessary. Simply activate swap to get rid of this problem.

starquake commented on 2022-01-04 08:27 (UTC)

My RPI 3B+ has problems compiling this package. It's running out of memory. Would you consider creating a binary package?

katoitalia commented on 2021-12-08 16:41 (UTC)

@max.bra yes, we all make mistakes but we all need to be corrected when we do. I am sorry if a simple feedback rocked your fragile self esteem, I can't help you with that.

max.bra commented on 2021-12-08 16:36 (UTC)

since you insist: then simply notify that the package has problems (which can happen, we are all human beings who can be wrong) without arrogating the right to teach me how to do it since you have NOT been asked. that's it. it is a matter of respect for the person and his work.

katoitalia commented on 2021-12-08 16:31 (UTC) (edited on 2021-12-08 16:35 (UTC) by katoitalia)

@max.bra please read it carefully, in no way shape or form that article on the wiki says that a broken package with an #outdated and broken# PKGBUILD shouldn't be considered as #outdated# and marked accordingly.

First, you should flag the package out-of-date indicating details on why the package is outdated, preferably including links to the release announcement or the new release tarball.

Should you find some other wiki article that proves me wrong, please quote it I will enjoy reading it. I am deeply sorry if feedback from the community isn't encouraged as it should.

To prevent such things to happen in the future consider checking if your packages do work before pushing them to the AUR.

Thank you.

max.bra commented on 2021-12-08 16:11 (UTC) (edited on 2021-12-08 16:15 (UTC) by max.bra)

... preferably including links to the release announcement or the new release tarball.
Nothing else to say or write.
p.s.: on forgetting on how to checksum a package: I was sarcastic :-D

katoitalia commented on 2021-12-08 15:09 (UTC) (edited on 2021-12-08 15:23 (UTC) by katoitalia)

@max.bra thank you for updating it quickly. It works just fine now. Cheers.

A broken PKGBUILD with outdated sha256sums is an outdated PKGBUILD.

Please consider building your packages locally with makepkg (and/or automating checks with a custom script) before pushing them to the AUR to avoid such issues.


max.bra commented on 2021-12-08 15:06 (UTC) (edited on 2021-12-08 15:13 (UTC) by max.bra)

@katoitalia hi, thanks for your explanation on a correct checksumming. after 5 years of packages maintaining i was forgetting it.
package updated with corrected sums. you can remove your post.
oh, and more: DO NOT flag package as outdated for a problem on it. simply posting the problem will notify me (and graysky) of everything.

katoitalia commented on 2021-12-08 14:41 (UTC) (edited on 2021-12-08 14:53 (UTC) by katoitalia)

lighttpd.pi-hole.conf fails sha256sum check
==> Validating source files with sha256sums...
    pi-hole-server-core-5.6.tar.gz ... Passed
    pi-hole-server-admin-5.8.tar.gz ... Passed
    arch-server-core-5.6-854957779.patch ... Passed
    arch-server-admin-5.8-854957779.patch ... Passed
    dnsmasq.include ... Passed
    lighttpd.pi-hole.conf ... FAILED
    nginx.pi-hole.conf ... Passed
    pi-hole.tmpfile ... Passed
    pi-hole-gravity.service ... Passed
    pi-hole-gravity.timer ... Passed
    pi-hole-logtruncate.service ... Passed
    pi-hole-logtruncate.timer ... Passed ... Passed ... Passed ... Passed

marking as outdated due to that

it should be


instead of

please update accordingly

Taijian commented on 2021-10-25 13:56 (UTC)

Thank you!

max.bra commented on 2021-10-25 13:45 (UTC)

hi Taijian, done w/o release bump

Taijian commented on 2021-10-25 13:38 (UTC)

Hi max.bra!

I'm aware of that option, but that would still leave this package dropping a permissions file on my system, granting a priviledge to a user that I do not want that user to have... Philosophically speaking, I find that offensive, because it makes the package willfully overwrite security decisions the sysadmin has made. I feel that this is exactly the kind of situation the the backup= array and .pacnew files are for.

So, could you please humour me? :)

max.bra commented on 2021-10-25 11:44 (UTC)

hi Taijian, there's no need to modify default file. you can create another file ('/etc/sudoers.d/my-pihole') containing 'pihole ALL=NOPASSWD: /usr/bin/pihole' to obtain same result and no overwrite risk.

Taijian commented on 2021-10-25 08:56 (UTC)

Hi @max.bra!

I have a tiny request to make: Could you please add /etc/sudoers.d/pihole to the backup= array? I'm asking because I run pihole's php-fpm instance as the pihole user instead of the http user, because there are other php-fpm instances also running on the same server, and so for security reasons I like to keep them seperate. I have therefore modified /etc/sudoers.d/pihole to accomodate that difference, but unfortunately the file gets overwritten every time pi-hole-server updates...

df8oe commented on 2021-10-03 08:01 (UTC) (edited on 2021-10-03 08:03 (UTC) by df8oe)

It can be done much easier...

1) pacman -S php7 php7-cgi php7-sqlite

2) open /etc/lighttpd/lighttpd.conf with a text editor and replace "/usr/bin/php-cgi" with "/usr/bin/php-cgi7", same with "/tmp/php-fastcgi.sock" which becomes "/tmp/php7-fastcgi.sock"

3) systemctl restart lighttpd.service

If you want to switch back to recent php just change the two text parts in lighttpd.conf (or hold a second configuration file which you apply at the start of lighttpd)

max.bra commented on 2021-10-02 07:00 (UTC)

@kusky thanks for reporting!

kusky commented on 2021-10-02 06:23 (UTC)

this package depends on core/inetutils /opt/pihole/ line 245: hostname: command not found

graysky commented on 2021-08-15 16:21 (UTC)

@max.bra - not sure what the problem was/didn't see anything diagnostic in logs, thanks for the offer. Will wait for next release and try that.

max.bra commented on 2021-08-15 14:58 (UTC)

@graysky sorry for late reply, back home today. i double checked my web admin interface and all seems normal... how can i help?

graysky commented on 2021-08-09 09:28 (UTC)

@max.bra - I downgraded to 5.3.1-3 when I found my web interface on 5.3.1-4 to be in a busy state. None of the menu items works and the banners across the top were displaying the animation for getting data. Pihole continued to process queries.

max.bra commented on 2021-06-28 10:46 (UTC)

@compiler1413 thanks for the heads up.
hopefully pihole 6 will be released soon and all of this problems will fade away.

compiler1413 commented on 2021-06-28 01:32 (UTC) (edited on 2021-06-28 01:39 (UTC) by compiler1413)

Just had hell of a time getting updated server to work properly. Arch Linux ARM, 5.11.2-1-ARCH #1 SMP PREEMPT Sun Feb 28 22:03:25 UTC 2021 armv7l

PHP8 is apparently not supported with the latest pi-hole-server so I downgraded to php7. This introduces an issue with SQLITE3.

lighttpd-angel[263]:mod_fastcgi.c.487) FastCGI-stderr:PHP Warning:  Use of undefined constant SQLITE3_OPEN_READONLY - assumed 'SQLITE3_OPEN_READONLY' (this will throw an Error in a future version of PHP) in /srv/http/pihole/admin/scripts/pi-hole/php/database.php on line 61
lighttpd-angel[263]:mod_fastcgi.c.487) FastCGI-stderr:PHP Fatal error:  Uncaught Error: Class 'SQLite3' not found in /srv/http/pihole/admin/scripts/pi-hole/php/database.php:42
lighttpd-angel[263]:mod_fastcgi.c.487) FastCGI-stderr:Stack trace:
lighttpd-angel[263]:mod_fastcgi.c.487) FastCGI-stderr:#0 /srv/http/pihole/admin/scripts/pi-hole/php/database.php(65): SQLite3_connect_try()
lighttpd-angel[263]:mod_fastcgi.c.487) FastCGI-stderr:#1 /srv/http/pihole/admin/api_db.php(56): SQLite3_connect()
lighttpd-angel[263]:mod_fastcgi.c.487) FastCGI-stderr:#2 {main}
lighttpd-angel[263]:mod_fastcgi.c.487) FastCGI-stderr:  thrown in /srv/http/pihole/admin/scripts/pi-hole/php/database.php on line 42

So I installed:

pacman -S php7 php7-cgi php7-sqlite php7-fpm

Recompiled pi-hole-server and pi-hole-ftl:

yay -S --aur pi-hole-server pi-hole-ftl

Made manual symlinks for php7 stuff:

ln -s /usr/bin/php-cgi7 /usr/bin/php-cgi
ln -s /usr/bin/php-config7 /usr/bin/php-config
ln -s /usr/bin/php7 /usr/bin/php
ln -s /usr/bin/phpize7 /usr/bin/phpize
ln -s /usr/bin/phar7 /usr/bin/phar
ln -s /usr/bin/php-fpm7 /usr/bin/php-fpm
ln -s /etc/php7 /etc/php
ln -s /usr/lib/php7 /usr/lib/php

Edited /etc/php/php.ini

sed -i "s|\;extension=pdo_sqlite|extension=pdo_sqlite|g;s|\;extension=sqlite3|extension=sqlite3|g;s|\;extension=sockets|extension=sockets|g" /etc/php/php.ini

Set immutable on php.ini because for some reason, after rebooting the settings were reverted

chattr +i /etc/php/php.ini

Restart lighttpd

sudo systemctl restart lighttpd

Everything is working as before now!

local/php7 7.4.20-1
local/php7-cgi 7.4.20-1
local/php7-fpm 7.4.20-1
local/php7-sqlite 7.4.20-1
local/pi-hole-ftl 5.8.1-1
local/pi-hole-server 5.3.1-3

And maybe I made an error or did something unnecessary here so please point that out for me if I did. Thanks!

pepper_chico commented on 2021-06-15 15:23 (UTC)

@max.bra, ok, thanks for the fix. I just thought that maybe other excerpts could need rebase reviewing, not only this special case. Maybe not.

max.bra commented on 2021-06-15 10:48 (UTC)

@pepper_chico oh! now i get it.
do you mean that our functions are not the same of upstream ones?
ok, it's true. thanks for reporting.
small oversight sorry, but this is not a job for anyone here. if you find a problem, report it and it will be fixed, if possible, ASAP.

max.bra commented on 2021-06-15 10:31 (UTC)

@pepper_chico i really don't understand

pepper_chico commented on 2021-06-15 10:24 (UTC) (edited on 2021-06-15 10:25 (UTC) by pepper_chico)

It seems forked sources here may need review & update. For example, I'm facing an issue that has been long fixed uptream, and should have been fixed here too, but isn't, because forked sources are not tracking upstream. See this fix present on 5.1.2 for example:

commented on 2021-05-09 00:26 (UTC)

Nevermind, deleting /etc/pihole/pihole-FTL.db and restarting the service fixed it.

commented on 2021-05-09 00:22 (UTC)

My long term data is not populating. Noticing some errors in the service when I run sudo systemctl status pihole-FTL.service:

● pihole-FTL.service - Pi-hole FTLDNS engine
     Loaded: loaded (/usr/lib/systemd/system/pihole-FTL.service; enabled; vendor preset: disabled)
     Active: active (running) since Fri 2021-04-16 11:13:00 EDT; 3 weeks 1 days ago
    Process: 352 ExecStartPre=/bin/chown -R pihole:pihole /etc/pihole (code=exited, status=0/SUCCESS)
    Process: 358 ExecStartPre=/bin/chown root:root /etc/pihole/logrotate (code=exited, status=0/SUCCESS)
   Main PID: 359 (pihole-FTL)
      Tasks: 7 (limit: 4915)
        CPU: 213ms
     CGroup: /system.slice/pihole-FTL.service
             └─359 /usr/bin/pihole-FTL no-daemon

Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.431 359M] Finished config file parsing
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.516 359M] Database version is 8
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] Updating long-term database to version 9
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] SQLite3 message: table aliasclient already exists in "CREATE TABLE aliasclient (id INTEGER PRIMARY KEY NOT NULL, name TEXT NOT NULL, comment TEXT);" (1)
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] ERROR: SQL query "CREATE TABLE aliasclient (id INTEGER PRIMARY KEY NOT NULL, name TEXT NOT NULL, comment TEXT);" failed: SQL logic error
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] ERROR: create_aliasclients_table() failed!
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] Aliasclients table not initialized, database not available
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] SQLite3 message: API call with invalid database connection pointer (21)
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] SQLite3 message: misuse at line 165242 of [5d4c65779d] (21)
Apr 16 11:13:00 alarmpi pihole-FTL[359]: [2021-04-16 11:13:00.517 359M] Error while trying to clos

steadybright commented on 2021-04-23 20:59 (UTC) (edited on 2021-04-23 21:29 (UTC) by steadybright)

I know this is getting off track, but I've since enabled IPv6 (within the last hour): All's well on the LAN side, but my ISP does not support IPv6, so I'm limited to to IP4 on the WAN, which is why I see the IPv6 connection errors, I suppose.

max.bra commented on 2021-04-23 19:52 (UTC)

I don't want to be intrusive: maybe ipv6 configuration was incomplete or only link-local (fe80::/10). mmmh, actually, specially with win10 clients a pi-hole protected lan should be fully ipv6 enabled.

steadybright commented on 2021-04-23 18:28 (UTC) (edited on 2021-04-23 19:22 (UTC) by steadybright)

@max.bra My pi-hole-server instance is running on my Arch Linux server. I have a Raspberry Pi 4B that also runs on Arch Linux (linux-raspberrypi, thanks to @graysky), but I do not run pi-hole on it. Some time ago, I disabled IPv6 on my Arch server because I was receiving incessant messages about connection failures from urbackup (among others) because they were looking to connect on IPv6. Disabling IPv6 fixed that issue, but apparently is causing other problems. It was a band-aid approach, and I suppose I should re-enable IPv6 on the server and just delete the logs more frequently.

max.bra commented on 2021-04-23 16:54 (UTC) (edited on 2021-04-23 17:02 (UTC) by max.bra)

you are welcome, thanks to you.
p.s.: obviously it would have been possible to solve by enabling a full ipv6 on your lan. if you are interested (choosing archlinux instead of a more classic raspberry distribution makes me think of it) would solve all your future problems without any system "hacking".

read: ... a more classic and pihole supported distribution ...

steadybright commented on 2021-04-23 16:43 (UTC)

That resolved it. Thank you again for you help with this--and for maintaining pi-hole-server for Arch!

max.bra commented on 2021-04-23 16:24 (UTC)

yes perfect. your pihole seems to be ok and ipv4 only. assuming your ISP is ipv6 compatible, i suspect that your win10 now relay (at it pleasure and after our tests) on ipv6 dns resolution. can you prioritize ipv4 following this '' (Solution 1 recommended)?

steadybright commented on 2021-04-23 16:05 (UTC)

@max.bra I haven't used pastebin before, but here's what I think you're looking for:

My browsing machine is Windows 10 and I'm using Firefox 88.0 64 bit

max.bra commented on 2021-04-23 15:55 (UTC) (edited on 2021-04-23 15:56 (UTC) by max.bra)

ok, step by step,
please: info about your browsing machine, in your pihole pastebin of 'pihole status' and a pastebin of your /etc/pihole/setupVars.conf

steadybright commented on 2021-04-23 15:52 (UTC) (edited on 2021-04-23 15:55 (UTC) by steadybright)

I just remoted in to another machine on my LAN and tried to navigate to a blocked site. It was blocked. Must be something to do with the machine I'm on physically. Thank you for your help!

steadybright commented on 2021-04-23 15:49 (UTC) (edited on 2021-04-23 15:49 (UTC) by steadybright)

I just re-confirmed 1. My router is using my pi-hole server for DNS, with no backup DNS server 2. Rebooted my router 3. Confirmed my local machine (from which I'm browsing) has my router as its DNS server, with no backup DNS server configured 4. Cleared my browser cache on my browsing machine 5. Rebooted my browsing machine 6. Confirmed that pi-hole-server is running, enabled, and without any errors. 7. Rebooted my pi-hole-server machine

Confirmed that nothing is being filtered. I don't know enough about DNS to know if there's a ghost somewhere causing this issue, but I've tried to run to ground all the obvious problems.

max.bra commented on 2021-04-23 15:27 (UTC)

@steadybright or maybe your desktop is on windows? ipconfig /flushdns?

max.bra commented on 2021-04-23 15:26 (UTC) (edited on 2021-04-23 15:30 (UTC) by max.bra)

@steadybright just tested: ookla speedtest.
pihole disabled, vivaldi no block: ads loaded
pihole enabled, vivaldi no block: ads NOT loaded

max.bra commented on 2021-04-23 15:23 (UTC)

@steadybright do you have any dns cache system in your machine? systemd dns stub?
or your browser do something you don't consider? for example: my vivaldi let me see ads only shutting down tracking service...

steadybright commented on 2021-04-23 15:16 (UTC) (edited on 2021-04-23 15:16 (UTC) by steadybright)

@max.bra thanks for the quick response. I updated to the latest (aur/pi-hole-server 5.3.1-2); however, now I can't seem to get pi-hole to filter. Everything is getting through. I confirmed that my DNS server is set to the pi-hole-server machine again. Re-booted the server and my local machine (from which I'm browsing). Could be my issue, but not sure what else to check. Oh, and I cleared my browser cache.

max.bra commented on 2021-04-23 14:41 (UTC) (edited on 2021-04-23 14:51 (UTC) by max.bra)

ok, found it. releasing ASAP.

max.bra commented on 2021-04-23 14:36 (UTC)

mmmh, there something wrong with 'pihole restartdns reload'...

max.bra commented on 2021-04-23 13:29 (UTC) (edited on 2021-04-23 14:04 (UTC) by max.bra)

@steadybright thanks for reporting. same here.
BLOCKING_ENABLED=true/false into /etc/pihole/setupVars.conf is correctly updated.
somebody else?

steadybright commented on 2021-04-22 15:13 (UTC) (edited on 2021-04-22 15:14 (UTC) by steadybright)

About a week ago, I noticed that when I disabled pi-hole via its web UI, pi-hole was still blocking sites as if it had not been disabled. There is no indication of failure in the web UI. It seems as if pi-hole believes it's disabled, but it is not. This happens whether I select Disable "Indefinitely," "For 5 minutes," "Custom time," etc.

Steps to reproduce the behavior:

  1. Go to 'http://<pihole_lan_ip>:808/admin/index.php'
  2. Login to the Web UI
  3. Click on 'Disable'
  4. Select any Disable option ("Indefinitely," "For 5 minutes," "Custom time," etc.)
  5. Observe "Enable(<COUNTDOWN>)" is active and counting down
  6. Try to open a blocked domain
  7. Observe that the domain is still blocked
  8. There is no error message. Pi-hole seems to think it's working as expected

max.bra commented on 2021-03-24 22:13 (UTC)

@jeffmikels no, the problem is in validIP($address) function that (in PHP8) will prevent insertion of v6_1 and v6_2 values at lines 181 and 184.
But believe me, with incoming pi-hole version 6, correcting it it's a complete waste of time.

jeffmikels commented on 2021-03-24 22:02 (UTC)

Making DNS Settings page compatible with PHP8...

in file /srv/http/pihole/admin/scripts/pi-hole/php/savesettings.php, replace two for loops...

details can be seen in my pull request here:

jeffmikels commented on 2021-03-24 21:51 (UTC)

php-fpm (8) behind nginx prevents access to the /etc/ directory entirely even when the web interface uses the pihole command line script!

My solution was to move /etc/pihole to /home/pihole and then create a symlink back to /etc/pihole.

Everything seems to be working from the web ui now.

max.bra commented on 2021-02-25 10:00 (UTC)

@all DNS servers settings page error: seems to be a PHP8 issue, if you need it and you can, please use PHP7

fl4co commented on 2021-02-08 14:01 (UTC) (edited on 2021-02-08 14:01 (UTC) by fl4co)

@max.bra that was quick, grazie!

max.bra commented on 2021-02-08 13:21 (UTC) (edited on 2021-02-08 13:21 (UTC) by max.bra)

1) if you need to set dns options via web interface you should
2) (unrelated note) thanks for reporting, try now.

fl4co commented on 2021-02-08 12:54 (UTC)

@max.bra yes, I'm ipv4 only. Looks like if I edited the php file the error would go away, but should I?

On an unrelated note, on two pi-holes I found out that if I disabled query logging via the web interface on the settings page, it's impossible to turn logging on again because the button doesn't turn into "enable query loging", as it does on the upstream pi-hole instead.

max.bra commented on 2021-02-07 21:17 (UTC)

@canton7 yes, already done elsewhere (see thanks for reporting, i'm adding that script...

max.bra commented on 2021-02-07 21:13 (UTC)

@fl4co php8 here and no problem. are you in a pure ipv4 scenario? (see

canton7 commented on 2021-02-07 13:15 (UTC)

I'm not sure where the responsibility for this issue lies, but I had problems with setting BLOCKINGMODE=IP-NODATA-AAAA.

The issue turned out to be that /srv/http/pihole/admin/scripts/pi-hole/php/queryads.php was trying to run sudo pihole ..., but /etc/sudoers.d/pihole only allowed accesss for /usr/bin/pihole, and not for a bare pihole.

I'm not sure what the best resolution here is: pihole.sudo from doesn't seem to contain any lines at all, so I'm not sure who's setting that up. Perhaps a patch on queryads.php to change the path to /usr/bin/pihole is the way to go?


fl4co commented on 2021-02-05 11:41 (UTC) (edited on 2021-02-05 11:41 (UTC) by fl4co)

Is it working with PHP 8? I'm seeing error messages on the web interface when changing upstream DNS servers.

pokexpert commented on 2021-01-17 10:32 (UTC)

@hmu yeah i already reported it. Quick dix dirty : replace the lsof from pihole script (there's only one) with

ss -lnp '( sport = 53 )'

It fixes the web interface

max.bra commented on 2021-01-17 10:14 (UTC)

@schmodd yes sorry, my fault. try now.

schmodd commented on 2021-01-17 10:11 (UTC) (edited on 2021-01-17 10:12 (UTC) by schmodd)

I'm getting with 5.2.3-1:

arch-server-core-5.2.3-764358184.patch ... FEHLGESCHLAGEN
arch-server-admin-5.3-764358184.patch ... FEHLGESCHLAGEN

hmu commented on 2021-01-11 21:21 (UTC)

Hi everyone, successfully running pihole with Apache since a few months, now after a recent update webinterface stopped working as well, taking ages to load. I think I have narrowed it down to the lsof -Pni:53 in the pihole script. It loads fine on the command line with sudo -u http /usr/sbin/lsof ..., so /etc/sudoers.d/pihole is working correctly, but apparently php running with apache just hangs when using exec('/usr/bin/lsof'); (tested in separate script). I'm not aware that I changed any security settings in either php.ini or httpd.conf, and my vhost definition basically just points to the admin/ dir. Is usage of exec('lsof'); restricted somewhere for apache? commented on 2021-01-04 21:43 (UTC)

@tkremer nice find. I've already uninstalled everything and went with installing this on a pi, which was my original plan anyways. I was just testing it out for a bit there on my manjaro installation. I hadn't been using lighttpd though either, I was/am using apache. I thought I had checked to make sure all of the apache and php settings were still correct, but I could have missed something for sure. Webservers are definitely not my area of expertise.

tkremer commented on 2021-01-04 14:33 (UTC) had the same problem and I can't really come up with an explanation but my /etc/php/php.ini was erroneous. It was missing the sqlite3 extension in the dynamic extensions section. After un-commenting the line and restarting lighttpd.service everything is working as expected again. commented on 2020-12-23 01:44 (UTC)

Latest version is broken for me. Web interface is not working. I'm running manjaro and apache and it had been configured and running for several months without any trouble. Latest update appears to have broken access to the web interfacee and in chronometer pi-hole is showing active, but FTL offline, and ads today and local qrys are 0%. Nothing under Top Advert/Domain/Client. It does seem to be blocking ads as far as I can tell though, I think. I don't know. I'm not sure what exactly to try and do to fix it at this point.

pw-sys commented on 2020-12-17 10:21 (UTC) (edited on 2020-12-17 10:22 (UTC) by pw-sys)

5.2.1 is broken for me with the webinterface running in apache. The problem is: I cannot access it at all, getting 408 or sometimes the login page shows after minutes but the login is failing timed out. Downgrading to my last installed version dbf2488c1ef4 solves the problem.

I have tried to access the interface from 3 different PCs 2 Android devices and via curl from the server itself. All have timed out.

rmi2 commented on 2020-12-15 20:11 (UTC)

I have a problem with the web interface. The statistics and settings are shown correctly, but no settings can be saved. No error messages are displayed (it even says that the settings have been updated; but they aren't).

  • I am using Nginx with php-fpm.
  • All packages are up to date as of today.
  • The ReadWritePaths are configured for php-fpm.
  • I do not have open_basedir enabled.
  • All needed services (nginx, php-fpm, pihole-ftl and pi-hole-gravity) have been restarted serveral times. I also tried "systemctl daemon-reload".
  • journalctl for nginx, php-fpm and pi-hole-gravity contain no error messages.
  • journalctl for pihole-ftl has a warning that I can't make sense of (and can't be googled): "WARNING: Unable to write PI"
  • In php-fpms journalctl, I can see the requests from the web interface: "Dez 15 21:03:55 xxx sudo[230857]: http : PWD=/srv/http/pihole/admin ; USER=root ; COMMAND=/usr/bin/pihole -a setdns,149.112>"
  • DNS works ("dig @" works). The requests are shown in the web interface.
  • "pihole status" says everything is fine.

Any ideas?

pokexpert commented on 2020-12-13 09:15 (UTC) (edited on 2020-12-13 09:18 (UTC) by pokexpert)

@setboolean i am running "true" arch. No manjaro or endeavour hère... I will try to recompile. Somthing to note is, everything works well in console, iincluding lsof...

xyzabc123 commented on 2020-12-13 08:51 (UTC) (edited on 2020-12-13 08:51 (UTC) by xyzabc123)

@pokexpert: 1. Are you running Arch or just a Arch flavored distribution? 2. Did you try to rebuild lsof from source? asp checkout lsof

max.bra commented on 2020-12-13 08:44 (UTC)

it would not be a problem at all to create a patch if it were a problem, I don't mean for everyone, but at least for the majority of users. if so, the developers would think about it directly. how many suffer from the problem on the execution of lsof?

pokexpert commented on 2020-12-12 22:38 (UTC)

@dkaea353 Thanks for the dirty fix ! Works well @max.bra it doesnt seems related to apache. Its indeed related to the lsof usage of the line 268 in pihole. Woudld it be too much to ask to create a .patch ? Or is it too dirty?

leo.sfsm commented on 2020-12-12 12:36 (UTC)

Shoudn't it be packaged on /usr/share/webapps? states that packages shouldn't write to /srv.

dkaea353 commented on 2020-12-09 20:52 (UTC)

Sorry, but right now I don't have the time to spare to look into this further. Just wanted to drop by and let you guys know a (very dirty) fix is replacing the "lsof -Pni:53" command in the /usr/bin/pihole script with the equivalent "ss -lnp '( sport = 53 )'". It's terrible, I know, but it's working...

max.bra commented on 2020-12-08 18:14 (UTC)

@pokexpert @dkaea353 i'm sorry but with "i've a problem.", nobody here can be of help. maybe a 'pastebin' of your system, arch type, package in use in your configuration (for example, pokexpert apache configuration is theoretically out of "support"), your own configuration files and all you can be able to communicate could make us understand what the problem may be.
i know, all pihole debug system has been removed and all must be done by hand: the price of "i file a bug upstream despite i'm using a non supported linux distro"...

dkaea353 commented on 2020-12-08 17:44 (UTC)

I can confirm the issue that pokexpert describes. "lsof -Pni:53" is being run as a result of "sudo /usr/bin/pihole status web" but is stuck at 100% CPU, which causes the page to load without end. Killing it then causes the page to be loaded with "DNS service not running" as status. Strangely enough, running "pihole status web" on the CLI just returns "1" as normal.

pokexpert commented on 2020-12-06 13:05 (UTC)

Ok so i got a weird issue for the web interface. I installed apache,php-sqlite and enabled sqlite extension. I did not do any apache config beyong enabling Php. When i open a page, it takes about a minute to answer. Page is fine. during that time, lsof -Pni:53 takes 100% cpu. Killing it makes the page appears with "dns service not running" on the status zone. Letting it load finally loads the page and "Active". This issue doesnt seems to be documented elsewhere. Something to ad,, if i go to pihole non admin "do you want to go to admin interface?", then every content gets 404 trying to go to pihole/pihole/admin/img/logo.svg (one pihole too much) This is annoying, could you help me?

xyzabc123 commented on 2020-12-05 22:05 (UTC)

@pato05 You're welcome.

pato05 commented on 2020-12-05 21:50 (UTC)

@setboolean Thank you.. I've just took a look in the wiki, and indeed saw I had to put these paths. I'm so sorry I asked such a dumb question, caused by the fact I did not read the wiki as stated in the pinned comment. Thank you again!

xyzabc123 commented on 2020-12-05 18:02 (UTC) (edited on 2020-12-05 18:03 (UTC) by xyzabc123)

@pato05: Did you add..

ReadWritePaths = /etc/pihole
ReadWritePaths = /srv/http/pihole php-fpm.service

pato05 commented on 2020-12-05 17:54 (UTC) (edited on 2020-12-05 17:55 (UTC) by pato05)

Hi! I installed it and did a bit of configuration, everything seems to be working and I can access the web interface (using nginx), but I can't actually edit any of the settings... I gave the http user all the needed permissions, but still, I can't change the settings... Any help?

wdiz commented on 2020-12-04 10:05 (UTC) (edited on 2020-12-04 10:05 (UTC) by wdiz)

@Taijian : Great it works! Sorry i didn't understand it was in the php-fpm.service file..My bad.

Taijian commented on 2020-12-04 09:23 (UTC)

@wdiz: In the end, I just did exactly what I wrote earlier: I simply removed the CapabilityBoundingSet= line in my php-fpm.service file. That fixed the issue for me, after first going through all the other steps that @max.bra kindly walked me through and not finding anything there.

wdiz commented on 2020-12-03 19:52 (UTC) (edited on 2020-12-03 19:53 (UTC) by wdiz)

@Taijian i do have the same issue but i can't manage to fix it. Can tell me more about files you modified ? Thx

max.bra commented on 2020-12-02 11:04 (UTC)

@Taijian glad you solved. see you next time. :-)

Taijian commented on 2020-12-02 11:02 (UTC)

@max.bra @setboolean: Thank you both for your help, I finally figured it out... Turns out, my php-fpm.service drop-in was too restrictive in the rights I was granting that service. It used to work with the restrictions in place, but now that it was trying to execute lsof with sudo, it no longer did because that got blocked by my CapabilityBoundingSet= settings. I commented that out and now it works again.

So, the important part for everyone reading this and the maintainers: The problem was entirely on my end, nothing to worry about for everyone else. Thank you again for the help!

xyzabc123 commented on 2020-12-02 10:52 (UTC) (edited on 2020-12-02 10:53 (UTC) by xyzabc123)

@Taijian @max.bra Hmmm strange, if pihole status and the API api.php?summaryRaw both report enabled the sudoers file should work as expected.

max.bra commented on 2020-12-02 10:50 (UTC) (edited on 2020-12-02 10:53 (UTC) by max.bra)

@Taijian or better: is your final part of your /etc/sudoers lihe this?

## Read drop-in files from /etc/sudoers.d
@includedir /etc/sudoers.d

edit: darn markdown!!

max.bra commented on 2020-12-02 10:46 (UTC)

@Taijian: no no no, http user cannot lsof directly, by web interface exec pihole script with sudo. is your sudoers.d/pihole over there?

Taijian commented on 2020-12-02 10:40 (UTC)

@max.bra: OK, I think your last comment is hitting on something. When I enter sudo -u http lsof -Pni:53 I get no output. sudo -u root lsof -Pni:53 on the other hand DOES give me something. Soooo, where did I go wrong, because I have to admit that this feels like it should tell me something, but I'm afraid that I'm too incompetent so see it...

max.bra commented on 2020-12-02 10:29 (UTC)

@Taijian: pihole script determines whether FTL is listening or not by this 'listening="$(lsof -Pni:53)"'. Is your http user blocked on this?

max.bra commented on 2020-12-02 10:24 (UTC)

@Taijian: anyway, if it's open_basedir related you must see something in your journal log...

max.bra commented on 2020-12-02 09:59 (UTC)

@Taijian: ok, that "led" is set by 'sudo pihole status web' that can return 0 or 1. does 'sudo pihole status' (w/o web) say same things?

Taijian commented on 2020-12-02 09:52 (UTC)

@max.bra and @setboolean:

Here's a link to the two relevant screenshots:

xyzabc123 commented on 2020-12-01 22:44 (UTC)

@Taijian: When you go to the following address: http://[IP_OF_PIHOLE]/admin/api.php?summaryRaw do you see "status":"enabled" or "status":"disabled"

max.bra commented on 2020-12-01 21:56 (UTC)

@Taijian can you shot me an image of your "red status button" please?

Taijian commented on 2020-12-01 21:50 (UTC)

@max.bra: Yes, I did (tried my backup Chromium, wiped that one's cache and tried again...). I do have to admit that I make use of open_basedir (and php-fpm is ofc also hardened), so it's not necessarily a widespread issue - but something must have changed from before, because now my setup has an issue...

I'm just trying to find out where that change might have occurred, so that I can adapt, not trying to complain, or anything!

max.bra commented on 2020-12-01 21:45 (UTC) (edited on 2020-12-01 21:46 (UTC) by max.bra)

@Taijian tried to wipe your browser cache (or use another one)?

Taijian commented on 2020-12-01 21:23 (UTC)

OK, I seem to have an issue with the latest update...

Pihole is undoubtedly running (resolves requests, reports as up on the CLI, bar graphs get generated on the web interface,...) however, the web interface shows a red status button and says that pihole is down. Any idea what changed from the previous version that I need to adjust my setup to?

max.bra commented on 2020-11-30 12:22 (UTC)

yes it is. thanks for reporting. added w/o rel bump

nandurius commented on 2020-11-30 12:19 (UTC)

It seems that lsof is now needed to correctly show the DNS resolver status both in pihole status and on the web interface. Would be great if you could add that as a dependency.

bitdruid commented on 2020-11-07 20:55 (UTC) (edited on 2020-11-07 20:57 (UTC) by bitdruid)

Hmm i have a problem with open_basedir. everything is configured just fine; Dashboard is loading but the settings tab always throws

....File(/var/log/apache2/error.log) is not within the allowed path(s): .... /srv/http/pihole/admin/scripts/pi-hole/php/auth.php:17

This path is in my open_basedir. anyone with the same bug?

max.bra commented on 2020-09-10 18:02 (UTC)

@ALoTron - yes yes, the problem is avoiding the error... maybe, in the end, on your system is worth putting /sys/class/thermal to openbase dirs to shut up php.

ALoTron commented on 2020-09-10 17:02 (UTC)

@max.bra Yes, there are a bunch of /sys/class/thermal/cooling_device* folders, probably one per core/thread as well. Pi-hole also displays a temperature, probably from /sys/class/hwmon/hwmon0/temp1_input (see /srv/http/pihole/admin/scripts/pi-hole/php/header.php:29). Getting a temperature is not the problem.

max.bra commented on 2020-09-10 16:38 (UTC) (edited on 2020-09-10 16:44 (UTC) by max.bra)

@ALoTron - this can explain the diversity. anyway, your system too must have a full path to term sensors in your thermal kernel module. please, try to find a full /sys/class/* or /sys/devices/* paths.

at this point the problem is not to have a symlinked path or not, but i think you have a totally different term thermal path to be read.

EDIT: on my workstation (ryzen too) under thermal i have a cooling_device* for every cpu core of my system...

ALoTron commented on 2020-09-10 16:02 (UTC)

@max.bra Well, I'm not using a Pi at all, Pi-hole runs on a desktop pc (Asus b350-Plus, Ryzen 1600x, 16GB RAM) with the 5.7.12-arch1-1 kernel.

max.bra commented on 2020-09-10 15:57 (UTC)

@ArchPCK - sorry read your edit only now. glad you solve. for your last question: live long and prosper :-D

max.bra commented on 2020-09-10 15:54 (UTC)

@ALoTron - oh!! on my Pi with archarm /sys/devices/virtual/thermal/thermal_zone0/temp is a fully functional path... what are your system and SO specs?

ALoTron commented on 2020-09-10 15:50 (UTC)

@max.bra But the same problem occures again, there is no /sys/devices/virtual/thermal/thermal_zone0/temp. The closes one is /sys/devices/virtual/thermal. Am I missing some thing on my system?

max.bra commented on 2020-09-10 15:44 (UTC) (edited on 2020-09-10 15:48 (UTC) by max.bra)

@ALoTron - mmh, /sys/class/thermal seems to be a little bit generic, opening the entire sensors subsystem. maybe is better to replace the symlinked path /sys/class/thermal/thermal_zone0/temp by /sys/devices/virtual/thermal/thermal_zone0/temp

ALoTron commented on 2020-09-10 15:30 (UTC)

@PS0 It seems like changing the path fixed it. Thanks for the help.

In addition to changing given path from /sys/class/thermal/thermal_zone0/temp to /sys/class/thermal Pi-Hole showed the same error message for /var/log/apache2/error.log. This path is not mentioned in the open_basedir section in the ArchWiki. Maybe it should be added to the rest?

ArchPCK commented on 2020-09-10 15:09 (UTC) (edited on 2020-09-10 15:16 (UTC) by ArchPCK)

@max.bra Yes I am using the config file provided by the package. I will try to find out more now

Edit: Ok I am so sorry, my setup script somehow failed to edit /etc/php/php.ini and did not uncomment the three modules... Sorry for wasting your time... And thanks for your support. Do you have a way I can donate to you?

max.bra commented on 2020-09-10 14:58 (UTC) (edited on 2020-09-10 14:58 (UTC) by max.bra)

@ArchPCK - seems only the php configuration in your lighttpd remains to be checked... followed pi-hole arch wiki? do you use lighttpd conf file provided with this package?

or better, is php functional in your web server? (your screenshot fail on first php content...)

ArchPCK commented on 2020-09-10 13:20 (UTC) (edited on 2020-09-10 13:22 (UTC) by ArchPCK)

@max.bra: yeah it shows the following:

[user@repolinux ~]$ sqlite3 /etc/pihole/gravity.db 'VACUUM;'
Error: attempt to write a readonly database

with sudo it shows no error, and I can rebuild it without any problems.

PS0 commented on 2020-09-10 09:18 (UTC)

@ALoTron I guess open_basedir doesn't deal with non-existing directories. Try changing it to the first existing parent directory (e.g. /sys/class/thermal/).

max.bra commented on 2020-09-10 07:44 (UTC) (edited on 2020-09-10 07:45 (UTC) by max.bra)

# sqlite3 /etc/pihole/gravity.db 'VACUUM;'

ArchPCK commented on 2020-09-10 00:11 (UTC) (edited on 2020-09-10 00:50 (UTC) by ArchPCK)

@max.bra: my pihole.log:

some screenshots:

No idea how to vacuum the gravity.db. I cant find anything in settings. Could you maybe give me an idea how to do it?

Updating the adlists in pihole web admin (update gravity) works just fine and it shows the output too.

Thanks for your time and work.

EDIT: it looks like its a lighttpd server error: But logs are emtpy. I will try to find out more after taking a nap.

ALoTron commented on 2020-09-09 20:19 (UTC)

@PS0 Nope, there is nothing at the given path, no symlink or file.

PS0 commented on 2020-09-09 14:10 (UTC)

@ALoTron: You have to put the real path(s) in open_basedir (without symlinks). The path /sys/class/thermal/thermal_zone0 is a symlink.

ALoTron commented on 2020-09-08 01:43 (UTC)

I have a problem with the open_basedir restriction. I added the paths given by the Archwiki but the web interface generates the following error message:

There was a problem applying your settings.
Debugging information:
PHP error (2): file_exists(): open_basedir restriction in effect. File(/sys/class/thermal/thermal_zone0/temp) is not within the allowed path(s): (/srv/http/pihole/:/run/pihole-ftl/pihole-FTL.port:/run/log/pihole/pihole.log:/run/log/pihole-ftl/pihole-FTL.log:/etc/pihole:/etc/hosts:/etc/hostname:/etc/dnsmasq.d/02-pihole-dhcp.conf:/etc/dnsmasq.d/03-pihole-wildcard.conf:/etc/dnsmasq.d/04-pihole-static-dhcp.conf:/proc/meminfo:/proc/cpuinfo:/sys/class/thermal/thermal_zone0/temp:/tmp:/sys/class/hwmon/hwmon0/temp1_input) in /srv/http/pihole/admin/scripts/pi-hole/php/header.php:25

For some reason it claims a path is missing, but in the same error message it shows this path in the open_basedir entry.

max.bra commented on 2020-09-07 16:22 (UTC)

@ArchPCK - oh, and of course, any errors in /run/log/pihole/.log?

max.bra commented on 2020-09-07 16:17 (UTC)

@ArchPCK - if you vacuum your gravity.db did you get any sort of error?

max.bra commented on 2020-09-07 15:27 (UTC)

@ArchPCK - can you 'ls -l' your /etc/pihole directory please?

ArchPCK commented on 2020-09-07 11:22 (UTC)

@graysky I tried opera, Firefox and chromium, all of them experience this

I did a fresh install and never opened anything on those browsers so cache was cleared.

Hope this gets solved somehow. It started with 5.x, 4.x was fine on that system.

graysky commented on 2020-09-07 10:31 (UTC)

@ArchPCK - Which browser? Cleared browser cache?

ArchPCK commented on 2020-09-07 04:15 (UTC) (edited on 2020-09-07 04:19 (UTC) by ArchPCK)

@wingsuit I have the same issue as you had, it is not a caching issue.

@max.bra do you have any ideas? In 4.3.1 and lower it was working fine with the guide on the wiki.

I can't add adlists, it just shows an "error has occurred", I get the same issue as you on the group site in the webinterface and my dashboard just contains the three boxes with a dash where the number should be, but no diagrams.

Upgrade existing packages and completely fresh os install is affected, I tried both. I thought about switching back to 4.3.1 but I can't compile that anymore.

However I was at least able to configure it by manually adding the adlists via CLI and editing the setupVars.conf file to change some of the settings..

But the web interface is still broken for me (tested using lighttpd). Only things working there is dhcp (but does not show clients) and DNS edit as well as restarting, flushing logs or changing log levels.

max.bra commented on 2020-08-26 14:30 (UTC)

@PS0 it's a good point for me. added w/o release bump.

PS0 commented on 2020-08-24 19:46 (UTC)

I know this has been discussed before, but sudo is missing as runtime dependency. The assumption that the base-devel group (with sudo in it) is installed only applies to build dependencies (makedepends and checkdepends), not to runtime dependencies (depends and optdepends). It is possible to build a package on one computer and install it on another computer (e.g. a server) which doesn't have base-devel installed. So, please add sudo to the 'depends' array.

Taijian commented on 2020-08-12 08:37 (UTC)

Reporting back: I have disabled pi-hole-logtruncate.timer on my server and instead copied /etc/pihole/logrotate to /etc/logrotate.d.

Works! No conflict, logs got rotated.

One thing though: In order to keep the old behaviour of daily log rotation, the entry for pihole-ftl.log needs to be changed from 'weekly' to 'daily'.

eh8 commented on 2020-08-12 02:18 (UTC)

@Taijian I didn't pick up on this until you mentioned. I have modified my pi-hole-logtruncate.timer file to run at a non-intersecting time with logrotate.timer

Taijian commented on 2020-08-11 10:28 (UTC) (edited on 2020-08-11 10:29 (UTC) by Taijian)

@max.bra & @grayski:

I am noticing a persistent issue on my pihole-server (which is also running various other services for me), namely that logrotate.service will regularly fail because of a concurrency issue with both logrotate.service and pi-hole-logtruncate.service attempting to execute the logrotate binary at the same time.

This got me thinking: Wouldn't it be a more 'clean' approach to scrap pi-hole-logtruncate.{service,timer} and instead add a proper config-snippet into /etc/logrotate.d/?

EDIT: speeeling is hrad!

k.w commented on 2020-08-10 09:01 (UTC)

@krupan yes, base-devel is needed for AUR. See

wingsuit commented on 2020-07-31 23:17 (UTC)

@graysky No tried on various cleared browsers. Seem to be having a lot of database related issues, many stats aren't showing up. Will try a fresh install.

graysky commented on 2020-07-31 12:41 (UTC)

@wingsuit - Browser cache? See recent discussion belo.

wingsuit commented on 2020-07-31 11:38 (UTC) (edited on 2020-07-31 11:41 (UTC) by wingsuit)

Dashboard completely empty after upgrade, no more graphs or anything, any ideas?

Also, when going to the Group tab I get "DataTables warning: table id=groupsTable - Invalid JSON response. For more information about this error, please see"

max.bra commented on 2020-07-19 07:01 (UTC)

@adityaruplaha all is funcional here, but i'm on lighttpd

adityaruplaha commented on 2020-07-19 04:04 (UTC) (edited on 2020-07-19 06:46 (UTC) by adityaruplaha)

Issue: "Unable to write to readonly database". This happens for group management, Local DNS Records, etc.

Configuration: Manjaro + Apache (PHP-FPM) + PiHole 5.1

Web server user http is in group pihole, and /etc/pihole/gravity.db has 'g+w' permissions.

I also asked on the PiHole forums:

krupan commented on 2020-07-18 19:29 (UTC) (edited on 2020-07-18 19:29 (UTC) by krupan)

I got:

==> ERROR: Cannot find the strip binary required for object file stripping.

Installing binutils fixed that. Then I got:

==> Starting prepare()...
/home/bmurdock/.cache/yay/pi-hole-ftl/PKGBUILD: line 38: patch: command not found
==> ERROR: A failure occurred in prepare().

Installing patch fixed that.

Then I got:

==> ERROR: Cannot find the fakeroot binary.

Installing fakeroot fixed that. Then I got another error at at this point I just installed all of base-devel.

eh8 commented on 2020-07-18 00:22 (UTC) (edited on 2020-07-18 02:26 (UTC) by eh8)

Post-install error, whenever you try to update DNS on the webpage you get:

/opt/pihole/ line 27: /etc/.pihole/automated install/ No such file or directory
/opt/pihole/ line 263: valid_ip: command not found
/opt/pihole/ line 263: valid_ip6: command not found

If you want a quick and dirty workaround, comment out line 27 and remove the conditional around line 263.

Noctem commented on 2020-07-17 22:29 (UTC)

I had to change some of the PHP invocations of pihole to /usr/bin/pihole for sudo to allow it and fix things like querying the block lists. I'm guessing this is related to PHP not expanding pihole to the full command path like a shell would and only /usr/bin/pihole being allowed in the sudoers file?

eh8 commented on 2020-07-16 05:54 (UTC)

I beg you to release v5.1 so I can gaze longingly at my bar graphs under the refreshing shade of the newly released dark theme.

max.bra commented on 2020-07-11 23:10 (UTC)

TheHardew commented on 2020-07-11 22:38 (UTC) (edited on 2020-07-11 22:38 (UTC) by TheHardew)

Please add patch to dependencies, failed in prepare without.

DanSman commented on 2020-06-20 23:10 (UTC) (edited on 2020-06-20 23:21 (UTC) by DanSman)

Can you guys add more lists/links to adlists.list? It doesn't get updated. It sticks to

after pihole -g



I've forgotten to use -r

pihole -g -r

deemon commented on 2020-06-19 22:35 (UTC)

what are those new configuration options and do we really need them?:


and why we don't need anymore?:


max.bra commented on 2020-06-19 06:26 (UTC)

hi ly50247, thanks for reporting!

ly50247 commented on 2020-06-19 02:10 (UTC)

dnsmasq.include should include addn-hosts=/etc/pihole/custom.list

samsa commented on 2020-06-07 11:30 (UTC) (edited on 2020-07-03 19:48 (UTC) by samsa)

I have problems with the permissions of the "/dev/shm/FTL-*" files. After reboot they are set to "root:root" and pihole works just for a short time and then stops resolving DNS queries with "Lost connection to API" on the web-UI. I have to manually chown the files to "pihole:pihole" and restart the service for the server to work reliably. Also all query data is lost every time I reboot the system or restart the pihole-FTL.service. Does anyone has a solution?

EDIT: Still have that problem. Can someone please tell me the standard permissions for /dev/shm/FTL-* ? I do not know if they have to be root:root. All problems are gone when I chown them to pihole:pihole and restart pihole-FTL.service.

Edit2: Can someone please tell me the output of "ls -l /dev/shm/FTL-*"?

erkexzcx commented on 2020-06-06 11:15 (UTC) (edited on 2020-06-06 12:54 (UTC) by erkexzcx)

Anyone has any idea why pi-hole is incredibly slow? takes like 15seconds to finish, when without pi-hole it takes like 1-2 sec?

EDIT: Pi-Hole f***ing sucks to me. I found alternative called grimd which works kinda same as Pi-Hole, without fancy charts, with nice config file and has super great performance while supporting DOH. I even made AUR package... Finally no more php/sqlite/web server nightmares and most important - no more issues with performance...

AlphaJack commented on 2020-05-30 14:19 (UTC)

Should pi-hole be moved under /usr/share/webapps instead of /srv/http/?

max.bra commented on 2020-05-25 09:13 (UTC)

@kikislater you are welcome!

kikislater commented on 2020-05-25 09:00 (UTC)

Oh well, understand ! I build it before upgrade

$ sudo pihole-FTL no-daemon
pihole-FTL: error while loading shared libraries: cannot open shared object file: No such file or directory

Rebuilt it and now ok, thank you for your support max and the pihole-FTL command

max.bra commented on 2020-05-25 07:34 (UTC)

hi iyedoost, yes it's enough.

iyedoost commented on 2020-05-25 04:57 (UTC) (edited on 2020-05-25 21:34 (UTC) by iyedoost)

Hi @max.bra, I'm using this package for updating my pihole gravity list, base on that i need to disable piholes auto update in this location: /etc/cron.d/pihole and this file doesn't exists and right now pihole update gravity list every week.

Edit: I just found out about pi-hole-gravity.timer and stop and disable it. Is that enough?

max.bra commented on 2020-05-24 17:04 (UTC)

@kikislater cay you please paste the result of

% sudo systemctl stop pihole-FTL
% sudo pihole-FTL no-daemon

kikislater commented on 2020-05-24 13:38 (UTC)

Hi, I don't manage to have FTL running fine ! I already stop and disbaled systemd-resolved as mentionned in wiki. any advises ?

$ sudo systemctl status pihole-FTL 
● pihole-FTL.service - Pi-hole FTLDNS engine
     Loaded: loaded (/usr/lib/systemd/system/pihole-FTL.service; enabled; vendor preset: disabled)
     Active: failed (Result: exit-code) since Sun 2020-05-24 17:33:49 +04; 2min 40s ago
    Process: 1253 ExecStartPre=/bin/chown -R pihole:pihole /etc/pihole (code=exited, status=0/SUCCESS)
    Process: 1254 ExecStartPre=/bin/chown root:root /etc/pihole/logrotate (code=exited, status=0/SUCCESS)
    Process: 1255 ExecStart=/usr/bin/pihole-FTL no-daemon (code=exited, status=127)
   Main PID: 1255 (code=exited, status=127)

May 24 17:33:49 vpsgiscan01 systemd[1]: pihole-FTL.service: Scheduled restart job, restart counter is at 5.
May 24 17:33:49 vpsgiscan01 systemd[1]: Stopped Pi-hole FTLDNS engine.
May 24 17:33:49 vpsgiscan01 systemd[1]: pihole-FTL.service: Start request repeated too quickly.
May 24 17:33:49 vpsgiscan01 systemd[1]: pihole-FTL.service: Failed with result 'exit-code'.
May 24 17:33:49 vpsgiscan01 systemd[1]: Failed to start Pi-hole FTLDNS engine.

graysky commented on 2020-05-22 17:22 (UTC)

@DocMAX - No, works fine in the the webUI here.

DocMAX commented on 2020-05-22 17:04 (UTC)

i cant disable pihole from web gui. its working on console with "pihole disable". anybody else?

Mouath commented on 2020-05-21 18:43 (UTC)

if you don't see charts in web interface, make sure you have sqlite enabled in /etc/php/php.ini

(mod_fastcgi.c.421) FastCGI-stderr: PHP Fatal error:  Uncaught Error: Class 'SQLite3' not found in /srv/http/pihole/admin/scripts/pi-hole/php/database.php:42
(mod_fastcgi.c.421) FastCGI-stderr: Stack trace:
(mod_fastcgi.c.421) FastCGI-stderr: #0 /srv/http/pihole/admin/scripts/pi-hole/php/database.php(65): SQLite3_connect_try()
(mod_fastcgi.c.421) FastCGI-stderr: #1 /srv/http/pihole/admin/scripts/pi-hole/php/gravity.php(13): SQLite3_connect()
(mod_fastcgi.c.421) FastCGI-stderr: #2 /srv/http/pihole/admin/index.php(68): gravity_last_update()
mod_fastcgi.c.421) FastCGI-stderr: #3 {main}
FastCGI-stderr:   thrown in /srv/http/pihole/admin/scripts/pi-hole/php/database.php on line 42

graysky commented on 2020-05-20 13:52 (UTC)

@max.bra - Yes, I believe clearing the cache did the trick.

max.bra commented on 2020-05-20 13:10 (UTC)

@graysky did you solve with the incorrect rendering of chromimum?

max.bra commented on 2020-05-19 18:26 (UTC)

@8633brown yes now i got it :-D

8633brown commented on 2020-05-19 18:17 (UTC)

@max-bra Sorry I dont thinkk i was very clear in my original comment...

this commit adds a new script for clearing the Arp/Network table using the command pi-hole arpflush. As this PKGBUILD currently stands this command currently fails as the is not copied into the "$pkgdir"/opt/pihole directory.

I'm not sure if I'm missing something or if this shouldnt be installed on Arch but as far as i can tell

install -Dm755 $_pkgname-$pkgver/advanced/Scripts/ "$pkgdir"/opt/pihole/

needs added to the PKGBUILD to stop the arpflush command failing.

ddifof commented on 2020-05-18 16:47 (UTC) (edited on 2020-05-18 16:51 (UTC) by ddifof)

I'm using this with Arch Linux ARM, I know it's not officially supported but just wanted to give some input on what I had to do to get things going after updating:

Do a Teleporter backup before updating!

  1. Pi-hole web interface

  2. Settings > Teleporter >Export

then update your base system, sudo pacman -Syu

After update pihole to 5.0 I had to do the following:

  • sudo pihole -g (to migrate and update gravity)

  • sudo pacman -S php-sqlite

Wasn't installed for some reason, I was getting errors like this from lighttpd:

PHP Warning: PHP Startup: Unable to load dynamic library 'pdo_sqlite' (tried: /usr/lib/php/modules/pdo_sqlite (/usr/lib/php/modules/pdo_sqlite: cannot open shared object file: No such file or directory), /usr/lib/php/modules/ (/usr/lib/php/modules/ cannot open shared object file: No such file or directory)) in Unknown on line 0

There was nothing rendered on the dashboard due to missing this, could not import teleporter backup either, and also explains why the Network and Long Term Data sections were non-working in 4.3.x for me.

  • Import your teleporter .tar.gz backup

  • Do another gravity update from the web interface

Hope this helps & thanks maintainers for the package.

max.bra commented on 2020-05-14 21:11 (UTC)

@8633brown the commit you linked do not remove the entire script, it removes lines from 39 to 45. mmmh, let's see what devs will do...

8633brown commented on 2020-05-14 20:05 (UTC)

v5.1 is missing the script

I'm not sure if any other files are missing.

recomended fix...

   install -Dm755 $_pkgname-$pkgver/advanced/Scripts/ "$pkgdir"/opt/pihole/
   install -Dm755 $_pkgname-$pkgver/advanced/Scripts/ "$pkgdir"/opt/pihole/
-  install -Dm755 $_pkgname-$pkgver/advanced/Scripts/ "$pkgdir"/opt/pihole/
   install -Dm755 $_pkgname-$pkgver/advanced/Scripts/ "$pkgdir"/opt/pihole/
   install -Dm755 $_pkgname-$pkgver/advanced/Scripts/ "$pkgdir"/opt/pihole/

graysky commented on 2020-05-13 01:24 (UTC)

@max.bra - Re: fonts, merged upstream,

Taijian commented on 2020-05-12 20:08 (UTC)

Here is an issue I have noticed - not sure if this is specific to a problem with my local config, or a general thing - so please do chime in...

1) Whenever I update my config via the 'Settings' tab of the web UI, setupVars.conf ends up being set to 600 - which then leads to the issue that the nginx user http cannot read it and an error message on accessing the tab again until I manually chmod 644 again.

2) A similar issue occurs with gravity.db. After I trigger a 'gravity update' via the web ui, gravity.db turns 644 instead of 664 that it needs to be in order for the nginx user http to be able to have write access to that db (which is necessary for editing block lists via the web UI).

So... Have I somehow borked my setup, or is this something that other nginx user also experience?

graysky commented on 2020-05-12 20:02 (UTC)

@max.bra - I will open an issue upstream. No need for +x on fonts.

max.bra commented on 2020-05-12 19:52 (UTC) (edited on 2020-05-12 19:59 (UTC) by max.bra)

@graysky wow!
1) they populate gravity.db during install script. we need a first run of pihole -g to migrate adlist.list to the new db during the first install of ver. 5.0+ package. maybe in the future we can do surely better.
2) COL_TABLE, non sense, a copy paste gift. corrected.
3) inherited from AdminLTE release. we only copy it.

francoism90 commented on 2020-05-12 19:05 (UTC) (edited on 2020-05-12 19:22 (UTC) by francoism90)

I really don't know if you should change it, but it's not conf.d for nginx but something like sites/pihole.conf instead.

Isproxy_intercept_errors on; really needed for pi-hole? I also don't think you need an extra /admin location.

Seems adding auth_basic would be useful as well:

graysky commented on 2020-05-12 18:25 (UTC) (edited on 2020-05-12 18:27 (UTC) by graysky)

@max.bra - I installed pihole in a container via their script just as a test. It seems that they do not provide adlist.list which we do:

Also, the following files are executable but I don't understand why they need to be:


max.bra commented on 2020-05-12 15:48 (UTC)

@graysky if pi.hole render bad and ip address not, i definitely bet on bad cache status

graysky commented on 2020-05-12 14:12 (UTC)

@max.bra - even more odd is that connecting to 'pi.hole' gives the weirdness but connecting to the ip address of the pihole box looks normal in chromium.

max.bra commented on 2020-05-12 14:00 (UTC)

firefox and chromium here too and the rendings are identical. maybe your chromium are chaching very bad for some strange reason?

graysky commented on 2020-05-12 13:56 (UTC)

@max.bra - I am using chromium and seeing the white circles and also a smooth plot of the queries. If I switch over to firefox, I get green circles and a block style plot for queries. So it seems to be a chromium problem???

max.bra commented on 2020-05-12 13:39 (UTC)

@graysky for example: for status icon this is the source

                    $pistatus = exec('sudo pihole status web');
                    if ($pistatus == "1") {
                        echo '<a id="status"><i class="fa fa-circle text-green-light"></i> Active</a>';
                    } elseif ($pistatus == "0") {
                        echo '<a id="status"><i class="fa fa-circle text-red"></i> Offline</a>';
                    } elseif ($pistatus == "-1") {
                        echo '<a id="status"><i class="fa fa-circle text-red"></i> DNS service not running</a>';
                    } else {
                        echo '<a id="status"><i class="fa fa-circle text-orange"></i> Unknown</a>';

as you can see pihole use font awesome and class text-'color' for the status. font awesome is ok, your browser does not honor the color class... is it possible? i don't think you are using internet exploder... :-D

graysky commented on 2020-05-12 12:56 (UTC)

@max.bra - any advice to troublshoot why my colors are white?

max.bra commented on 2020-05-12 12:41 (UTC) (edited on 2020-05-12 12:42 (UTC) by max.bra)

Hi graysky! Nope, colors here:
about logrotate: always been there. pihole scripts are calling it in that path... it's not really a standard logrotate...
edit: wording

graysky commented on 2020-05-12 12:28 (UTC)

@max.bra - Thanks for the update. After I updated to the new versions, I ran pihole -g and restarted lighttpd. Everything seems to be working. Two observations:

1) pi-hole-server owns /etc/pihole/logrotate but should it be in that dir? 2) Are the buttons under "status" in the upper left white for you too?

erion commented on 2020-05-12 12:03 (UTC)

@Taijian You are right, I did miss that step somehow. My guess is that the php-fpm instructions are under a Nginx section, so I skipped past it.

Creating the service override does the trick and now I am back to about 950000 blocked domains.

Thanks for pointing me in the right direction, as well as thanks to everyone who replied, making sure that I was doing the steps properly :)

Taijian commented on 2020-05-12 10:49 (UTC)

@erion: Have you also modified php-fpm.service as mentioned in the linked wiki article? I'm asking because you specifically mention that you have done all the other recommended steps, but left that one out...

erion commented on 2020-05-12 08:00 (UTC)

I mentioned in my comment that I am using Caddy, I am sorry if this was unclear.

I was referring to directives in /etc/php/php.ini.

Caddy's fastcgi directive can pass environment variables, if needed, but it shouldn't touch open_basedir, which means that it should honor whatever's in php.ini if I am not mistaken. Am I misunderstanding something?

max.bra commented on 2020-05-12 07:49 (UTC)

or an equivalent in your nginx.conf

max.bra commented on 2020-05-12 07:47 (UTC) (edited on 2020-05-12 07:47 (UTC) by max.bra)

@erion no way... then you have

  -d open_basedir=....

in your lighttpd.conf, last chance

erion commented on 2020-05-12 07:07 (UTC)

@max.bra Yes, at the moment it is commented. I tried with and without a semicolon.

max.bra commented on 2020-05-12 06:43 (UTC)

@erion you mean "COMMENTED the open_basedir directive" right?

i wrote:

; open_basedir =

note: semicolor

erion commented on 2020-05-12 04:36 (UTC)

Yes, I have. Sqlite and pdo_sqlite are enabled, as well as sockets, and I have uncommented the open_basedir directive. It has an empty value.

k.w commented on 2020-05-11 20:19 (UTC)

@erion did you see the new required config for php-fpm since 7.4?

max.bra commented on 2020-05-11 19:34 (UTC)

@Mettacrawer Does this package seem pi-hole-ftl to you?

Mettacrawer commented on 2020-05-11 18:57 (UTC)

/usr/lib/ was upgraded to /usr/lib/

/usr/bin/pihole-FTL was linked against /usr/lib/ so it would not start until I rebuilt the pi-hole-server package with yay -S --rebuild pi-hole-ftl

Simply running yay -Suy was not good enough.

Please add dependencies for the shared libraries that this package uses. Thanks

erion commented on 2020-05-11 18:25 (UTC)

@max.bra No dice. I've done the following:

  1. *gpasswd -a http pihole
  2. reboot.
  3. Restart web browser.
  4. At this point, groups http correctly outputs: pihole http
  5. After logging in and trying to add an adlist, I am still getting the "unable to write a read-only database" error.

max.bra commented on 2020-05-11 18:05 (UTC)

@iyedoost -r pihole option is not present into this package

@erion assumed you are using lighttpd or nginx, add http user to pihole group with method you prefer, then restart your web browser and try again.

erion commented on 2020-05-11 18:02 (UTC)

@iyedoost As far as I know, this is executed when pihole -g is ran for the first time for v5. Nonetheless, I tried, unfortunately no change.

erion commented on 2020-05-11 17:56 (UTC)

Sorry, I meant groups http.

Unfortunately, I did not see pihole in its output, even after a restart.

Testing it right now, and for some reason typing groups on its own, then groups http now lists pihole as expected. Unfortunately, the read-only database issue is not solved by this, so my scenario must be different compared to what the upstream devs had suggested for someone else.

iyedoost commented on 2020-05-11 17:55 (UTC)

@erion try 'pihole -g -r' in command line

max.bra commented on 2020-05-11 17:47 (UTC)

"groups http" will return http and pihole. groups pihole not.

restart your web server or reboot.

erion commented on 2020-05-11 17:43 (UTC)

After installing the 5.0 update, I am trying to add a new adlist, but I get:

Error: Something went wrong while executing: Attempt to write a read-only database.

/etc/pihole is owned by the user pihole. My web server (Caddy) runs under http, which is added to the pihole group.

Interestingly, when I use gpasswd or usermod to add http to the pihole group, and execute the groups pihole command afterwards, I don't see http listed, however, deleting the user from the group works as expected.

Any ideas about what could be wrong?

iyedoost commented on 2020-05-11 16:32 (UTC) (edited on 2020-05-11 16:33 (UTC) by iyedoost)

@max.bra You're AWESOME That fix the problem. Thank you ;)

max.bra commented on 2020-05-11 16:21 (UTC)

@iyedoost open_basedir has nothing to do with web servers, it's a configuration of your php. please edit your /etc/php/php.ini and unset your open_basedir directive:

;open_basedir =

note: since this package has never edited php.ini, you have edited and modified your open_basedir directive.

iyedoost commented on 2020-05-11 16:13 (UTC)

max.bra I add that dir (apache log) to php.ini and I'm still see this error. and pihole still don't have any adlist:

you think it might be lighttpd problem? and i shoud replace it with apache?

Taijian commented on 2020-05-11 15:36 (UTC)

suggestion for v5: Maybe you could add an optdepend on php-sqlite?

People who are using this with php-fpm via nginx/apache will otherwise be unable to access the database settings from the web interface.

iyedoost commented on 2020-05-11 15:32 (UTC)

@max.bra I'm using lighttpd and in wiki page "/var/log/apache2/error.log" is not in the list.

right now pi-hole ver 5 block nothing at all and i cant edit adlist. see this:

iyedoost commented on 2020-05-11 15:19 (UTC)

and i can't add adlist i get this error: Error, something went wrong! While preparing statement: no such table: adlist


iyedoost commented on 2020-05-11 14:50 (UTC)

i see this error in settings page

im not using apache i use lighttpd

max.bra commented on 2020-05-11 14:05 (UTC)

ver. 5.0 note: after updating please remember to restart your web server or simply reboot.

max.bra commented on 2020-05-09 19:31 (UTC) (edited on 2020-05-09 19:31 (UTC) by max.bra)

for the flag enthusiasts:

1) eventually, the right package to flag is FTL
2) ftl is not outdated, it needs only to be recompiled

dlins commented on 2020-05-08 21:33 (UTC) (edited on 2020-05-08 21:35 (UTC) by dlins)

It is broken. The binary needs, but arch provides

@max.bra Sry didn't see your comment.

max.bra commented on 2020-05-08 09:35 (UTC)

This is not outdated, it need recompilation.

lewis2e commented on 2020-03-30 15:13 (UTC)

@max.bra Oh yeah! Sorry. I misread your comment! Cheers, anyway.

max.bra commented on 2020-03-30 15:10 (UTC)

glad you solve. anyway i was not speaking about /etc/resolv.conf... :-D

lewis2e commented on 2020-03-30 15:06 (UTC) (edited on 2020-03-30 15:15 (UTC) by lewis2e)

@max.bra Your suggestion gave me enough information to solve this! Massive thank you.

If you (or anyone) is interested...

I was unable to add the DNSStubListener=no option to /etc/resolv.conf as suggested* because I am using openresolv to manage that file. This is required by netctl which I am using as my network manager.

*EDIT - I misread @max.bra's comment. This was exactly what they were suggesting anyway :-)

I am sure that you probably can get openresolv to add the option, but I couldn't figure out how.

The solution was to instead add DNSStubListener=no to /etc/systemd/resolved.conf. This FINALLY seems to stop the dam thing binding to port 53 on boot, and pihole starts just fine. I note that systemd-resolved is still running (despite me disabling it), but seems to be keeping to itself now. There really is no killing that thing apparently....

Thank you for the help. Nice to get rid of this little paper cut every time I reboot my server.

max.bra commented on 2020-03-30 14:10 (UTC)

but... did you


in your /etc/systemd/resolved.conf?

max.bra commented on 2020-03-30 14:07 (UTC)

What about:

# grep -R -i resolved /etc/systemd/system/*
# grep -R -i resolved /lib/systemd/system/*

lewis2e commented on 2020-03-30 14:04 (UTC) (edited on 2020-03-30 14:05 (UTC) by lewis2e)

@max.bra I don't recall doing so, however, I built the system quite a long time ago now. I could not swear to it. More to the point, even if I did enable it, I should be able to disable it in the usual way (which I cannot). If the implication is that I am not using Arch, I am using Arch Linux Arm.

max.bra commented on 2020-03-30 13:56 (UTC)

@lewis2e in arch systemd-resolved it is not automatically enabled. that said, it should have been you who configured and enabled it.

lewis2e commented on 2020-03-30 10:25 (UTC)

Following every reboot, pihole.FTL fails to start. I have to manually stop the systemd-resolved.service and then restart pihole.FTL or it won't work.

I've tried disabling the systemd-resolved.service as described in the Arch Wiki, but it still starts up. I don't know how to find out what is calling for it. Any suggestions?

max.bra commented on 2020-03-23 10:27 (UTC)


ilpianista commented on 2020-03-23 10:19 (UTC)

Please note that sudo is required when using the UI:

lighttpd-angel[11938]: sh: sudo: command not found

DanSman commented on 2020-02-25 21:53 (UTC) (edited on 2020-02-25 22:04 (UTC) by DanSman)


That worked for me. Thanks!

But I only added "ReadWritePaths" in the php-fm.service. That was enough.

I have 2 piholes running and it must changed after the php update to 7.4, because it worked before. I'm going to move/merge my old server to the new server. So one pihole installation is quite new. The old one is with openbasedir and the new one is without. Was the same problem.

'Network overview' isn't working on the new server. I don't care at the moment ;)

There is also a permission problem when I flush the logs. At least in the terminal:

sudo pihole -f
  [i] Flushing /run/log/pihole/pihole.log ...error: Ignoring /etc/pihole/logrotate because the file owner is wrong (should be root or user with uid 0).
error: Ignoring /etc/pihole/logrotate because the file owner is wrong (should be root or user with uid 0).
  [✓] Restarting DNS service
  [✓] Flushed /run/log/pihole/pihole.log
  [✓] Deleted 149518 queries from database

I can change it to root but every restart of pihole will change it anyway.

xyzabc123 commented on 2020-02-25 08:50 (UTC)

@DanSman: Had the same problem and fixed it by adding the pihole http and /etc/pihole directory to "ReadWritePaths" in the php-fm.service file via drop-in.

max.bra commented on 2020-02-25 08:11 (UTC)

@beedaddy mmm, yes maybe...

but I'm on php 7.4 on my PI and no problem here...

beedaddy commented on 2020-02-25 08:00 (UTC)

My guess is that DanSman experienced the same problem I did (see my comment from 2019-12-03 07:52). The solution was to downgrade from php 7.4 to 7.3.

max.bra commented on 2020-02-25 07:13 (UTC)

Hi DanSman, are you in a fresh install scenario?
Are logs saying something usefull?

The last 3 updates deal only with internal installation procedures which we do not actually use. It is therefore (nearly) impossible that the last package could have caused you such problems.

Did you follow the wiki in the web interface section, especially the one about PHP openbasedir (if you use it)?

DanSman commented on 2020-02-25 05:19 (UTC)

Hi max.bra

Thanks for pihole on Arch.

I'm trying to make some changes in setting, whitelist, blacklist etc. but for some reason nothing changed. I have to change it through terminal.

Tried already some permissions changes but still the same problem.

Need a little help here.

Thanks Dan

beedaddy commented on 2019-12-03 07:52 (UTC) (edited on 2019-12-04 14:20 (UTC) by beedaddy)

I have a problem with a fresh installation on AArch64 using nginx as web server. The status on the admin page is "unknown" and it is not possible to save changes in settings. I set it up step by step according to the pi-hole Arch Wiki page.

The pihole-FTL service starts without problems. Also, the http user is fine (not in expired state).

The only error I noticed in journal log is:

[error] 448#448: *2498 FastCGI sent in stderr: "PHP message: PHP Notice  Trying to access array offset on value of type null /srv/http/pihole/admin/settings.php on line 23PHP message: PHP Notice:  Trying to access array offset on value of type null in /srv/http/pihole/admin/settings.php on line 23" while reading upstream, client:, server: _, request: "GET /admin/settings.php HTTP/1.1", upstream: "fastcgi://", host: "", referrer: ""

Perhaps it's a problem with php 7.4?

EDIT: Indeed, downgrading to php 7.3 resolves the issue.

max.bra commented on 2019-11-19 10:06 (UTC)

so, assuming you have also done this ( and this ( you should only need to restart lighttpd...

DNAblue2112 commented on 2019-11-19 10:01 (UTC) (edited on 2019-11-19 10:03 (UTC) by DNAblue2112)

yes, I have installed and reinstalled them both in my troubleshooting efforts

EDIT: it seems to have started working, I was just restarting services and something must have reloaded its config or something because its all working fine now.

max.bra commented on 2019-11-19 09:50 (UTC)

mmm, are sqlite AND php-sqlite installed?

DNAblue2112 commented on 2019-11-19 09:44 (UTC) (edited on 2019-11-19 09:48 (UTC) by DNAblue2112)

I checked the permissions, they all look good.

It seems that it is an SQLite issue, I have followed the guide at yet I still found this error

Nov 19 20:11:36 server lighttpd-angel[268]: 2019-11-19 20:11:36: (mod_fastcgi.c.421) FastCGI-stderr: PHP Fatal error: Uncaught Error: Class 'SQLite3' not found in /srv/http/pihole/admin/api_db.php:71

Nov 19 20:11:36 server lighttpd-angel[268]: 2019-11-19 20:11:36: (mod_fastcgi.c.421) FastCGI-stderr: Stack trace:

Nov 19 20:11:36 server lighttpd-angel[268]: 2019-11-19 20:11:36: (mod_fastcgi.c.421) FastCGI-stderr: #0 /srv/http/pihole/admin/api_db.php(86): SQLite3_connect(true)

Nov 19 20:11:36 server lighttpd-angel[268]: 2019-11-19 20:11:36: (mod_fastcgi.c.421) FastCGI-stderr: #1 {main}

Nov 19 20:11:36 server lighttpd-angel[268]: 2019-11-19 20:11:36: (mod_fastcgi.c.421) FastCGI-stderr: thrown in /srv/http/pihole/admin/api_db.php on line 71

max.bra commented on 2019-11-19 09:38 (UTC) (edited on 2019-11-19 09:38 (UTC) by max.bra)

hi DNAblue2112, seems a permission problem. check /etc/pihole/* for files that are not owned by pihole user.
'journalctl -b' should provide any useful information.

DNAblue2112 commented on 2019-11-19 09:27 (UTC)

I receive an error whenever I attempt to access the networks tab or the long term data query log. Both errors are because of a 500 response from api_db.php In chrome console the error is as follows: api_db.php?network&_=1574155379457:1 Failed to load resource: the server responded with a status of 500 (Internal Server Error) I have made sure that SQLite is installed and enabled in PHP. Most functionality works so its not a massive problem but it would be nice to be able to use all of the features. What other error logs can I provide to help resolve this?

denisse commented on 2019-10-16 21:57 (UTC)

Thank you so much @Prakkie and @max.bra, the problem is solved now!

max.bra commented on 2019-10-16 15:32 (UTC)

hi alephZer0, please check the first pinned comment on this page.

If it's not your case, can you pastebin (or any other service) this?

    ls -l /etc/pihole

Prakkie commented on 2019-10-16 15:29 (UTC)

@alephZero, did you happen to check pinned comment and try?

Checking /etc/shadow, the user http had a trailing 1:

$ sudo cat /etc/shadow | grep http http:!!:18174::::::1: Removing the expiration date has resolved it:

$ sudo chage --expiredate -1 http

max.bra commented on 2019-10-16 15:28 (UTC) (edited on 2019-10-17 06:18 (UTC) by max.bra)

Since switching to pihole user for pihole-FTL - long term data is not retained on my machine.

sir_randomuser said: For some reason account http has expired, which was the root cause. Seems to be due to a recent systemd change:
Checking /etc/shadow, the user http had a trailing 1:

$ sudo cat /etc/shadow | grep http

Removing the expiration date has resolved it:

$ sudo chage --expiredate -1 http

denisse commented on 2019-10-16 15:19 (UTC)

I'm sadly unable to save settings for my Pi Hole using the web interface.

The web interface gives me a message saying that everything was saved buy when I reload, nothing was saved.

The Pi Hole is working in the sense that it's blocking ads, but I want to use it a my DHCP server but I'm unable to do it as the setting is not saved.

Could you please give me a hand?

Prakkie commented on 2019-10-15 22:35 (UTC)

wonderful @max.bra. It works now. GREAT THANKS.

max.bra commented on 2019-10-15 21:50 (UTC)

hi Prakkie, read the first pinned comment on page.

Prakkie commented on 2019-10-15 21:05 (UTC)

pihole on arch - webui shows status as unknown

After upgrading to latest arch with linux version 5.3.6, pihole web UI comes with status as "Unknown" and "Enable" button. When i click on enable button, status becomes "Active" however its not persistent. The moment i click on any other option in webui, status goes back to "Unknown".

I am not sure what i am missing. pihole -d is not available in arch. CLI shows pihole is active. I am not finding any errors in nginx logs too.

I followed instructions in arch pihole page. I tried both nginx as well as lighttpd.

graysky commented on 2019-09-25 20:46 (UTC) (edited on 2019-09-25 20:47 (UTC) by graysky)

I think the issue is that @max.bra has created the same name of the patches but has modified them since doing so. Therefore, if a user built an older version expecting the patch, the checksums would be different. One solution is to append a date to the patch itself in the source array as suggested below, another is to use use a tag in the github repo.

git tag $tag -m "$tag" && git push --tags

jshap commented on 2019-09-25 20:12 (UTC) (edited on 2019-09-26 00:49 (UTC) by jshap)

it's not a yay issue, it happens with just makepkg as well if you don't delete the packages first. it's just a normal issue with how makepkg chooses to download things based on if it thinks it's changed or not.

for a more esoteric use case, look at how firefox-nightly's pkgbuild works. for them the url does not change, and yet every day the tarball will. so in the source array they move the file to a different location based on $(date ...). obviously not asking you to do that, just saying that it is not completely unheard of to have to rename things which are downloaded.

but again, I'm not saying to attach the date, I'm saying just use the pkgrel.

tholinka commented on 2019-09-25 05:19 (UTC)

Deleting the patches (or just clean building) does fix the checksum issue.

However, an even better solution (imo), since you already have to update this package with the new checksums, is to just copy the patches into this package instead of having them source from github. This comes with the added benefit that it's easy to see a diff between the versions.

max.bra commented on 2019-09-24 19:58 (UTC)

and more:

[Aur: 1]  pi-hole-server-4.3.2-5

  1 pi-hole-server                   (Build Files Exist)
==> Packages to cleanBuild?
==> [N]one [A]ll [Ab]ort [I]nstalled [No]tInstalled or (1 2 3, 1-3, ^4)

this appear a clear question to me. or not? it's just a matter of not pressing <enter> blindly...

max.bra commented on 2019-09-24 19:53 (UTC)

@all sorry but, i need to change the PKGBUILD because of yay build caching? am i understanding well? this is not normal and the use of an aur helper is not mandatory.

cat .config/yay/config.json 
"buildDir": "/tmp",
"editor": "",

jshap commented on 2019-09-24 19:02 (UTC) (edited on 2019-09-24 19:05 (UTC) by jshap)

the issue isn't that the checksums are wrong it's that arch-server-admin-4.3.2.patch and arch-server-core-4.3.2.patch change with each pkgrel bump but their url doesn't, so although the server is different than your downloaded version because they share the same name they don't get updated, and then fail the checks.

solutions would be to update to name the patches like arch-server-core-4.3.2-5.patch, or to update the PKGBUILD to download them and save them by pkgrel, so the sources would look like:



prepare() {
  cd "$srcdir"/"$_pkgname"-"$pkgver"
  patch -Np1 -i "$srcdir"/arch-server-core-$pkgver-$pkgrel.patch
  cd "$srcdir"/"$_wwwpkgname"-"$_wwwpkgver"
  patch -Np1 -i "$srcdir"/arch-server-admin-$_wwwpkgver-$pkgrel.patch

johnnybash commented on 2019-09-24 18:18 (UTC)

or just configure your AUR helper to either cleanbuild (as stated several times) or your build directory into a temp dir, yay e.g. supports both... -.-

oldrava commented on 2019-09-24 18:16 (UTC) (edited on 2019-09-24 18:17 (UTC) by oldrava)

Current easy workaround(for yay) yay -G pi-hole-server , cd pi-hole-server , makepkg -g >> PKGBUILD , makepkg , sudo pacman -U pi-hole-server-4.3.2-5-any.pkg.tar.xz

pi-hole-server-4.3.2-5-any.pkg.tar.xz is current version change it accordingly to it version and file it created

oldrava commented on 2019-09-24 18:09 (UTC)

Dear max.bra

I appreciate your work but if I get ==> ERROR: One or more files did not pass the validity check! there is no use of your repo if I had to download and compile it myself, Kind regards

max.bra commented on 2019-09-24 15:31 (UTC) (edited on 2019-09-24 15:43 (UTC) by max.bra)

Hi gecko, dns servers list and adblocks list are taken manually from upstream It happens that something can be lost especially for the informations extrapolated from precisely. In fact, the AUR community never fails to notice what others do. And this is very good.

Given that pi-hole on archlinux is not supported by the original developers and the use of AUR packages is at the user's own risk, the update details are all opensource. You can find all the installation files here in the git repository of this package (FTL details on the respective page) and the patch generator on the respective github repos (

Updating adblock list... done!! Thanks for reporting, as usual. ;-)

gecko commented on 2019-09-24 15:02 (UTC) (edited on 2019-09-24 15:15 (UTC) by gecko)


Firstly, let me preface by saying thank you for the work you do on maintaining this aur package. It’s great work you’re doing and you make it very convenient for a lot of people, so I’m sure I’m not alone in saying thank you - we appreciate your effort and hard work :) May I ask, how does your package work exactly when updates are applied upstream? Because I noticed in the pihole 4.3.2 release notes ( that the zeustracker blocklist has been removed from the default blocklists contained in adlists.list file. This was to fix PR#2843 (

However, upon installing your package fresh (ie not an upgrade, full uninstall and removal of /etc/pihole/ folder then reinstall) I find that the zeustracker is still contained in the default blocklist adlists.list file. I assumed that all changes made upstream are automatically applied to this package when there is an update, but this appears to not be the case. So I was wondering if you would mind going into some detail as to what happens when you update this package to reflect an upstream update?

Thanks again for your work :)

xyzabc123 commented on 2019-09-23 09:50 (UTC) (edited on 2019-09-23 09:50 (UTC) by xyzabc123)

@nealc99 try a clean build

nealc99 commented on 2019-09-23 09:26 (UTC) (edited on 2019-09-23 09:28 (UTC) by nealc99)

getting checksum error for arch-server-core-4.3.2.patch
==> Validating source files with md5sums... pihole-4.3.2.tar.gz ... Passed admin-4.3.2.tar.gz ... Passed arch-server-core-4.3.2.patch ... FAILED arch-server-admin-4.3.2.patch ... Passed dnsmasq.include ... Passed lighttpd.pi-hole.conf ... Passed

max.bra commented on 2019-09-21 10:50 (UTC)

@phixion YES!! really long forgotten... thanks for reporting

phixion commented on 2019-09-21 03:32 (UTC)

shouldn't version tag in and line 22 be raised to 4.3.2 ?

johnnybash commented on 2019-09-19 05:08 (UTC)

as have to cleanbuild...

MagicBobert commented on 2019-09-18 23:35 (UTC)

The md5sum for arch-server-admin-4.3.patch is not correct.

==> Validating source files with md5sums...
    pihole-4.3.2.tar.gz ... Passed
    admin-4.3.tar.gz ... Passed
    arch-server-core-4.3.2.patch ... Passed
    arch-server-admin-4.3.patch ... FAILED
    dnsmasq.include ... Passed
    lighttpd.pi-hole.conf ... Passed
    nginx.pi-hole.conf ... Passed
    pi-hole.tmpfile ... Passed
    pi-hole-gravity.service ... Passed
    pi-hole-gravity.timer ... Passed
    pi-hole-logtruncate.service ... Passed
    pi-hole-logtruncate.timer ... Passed ... Passed
    version.patch ... Passed ... Passed
==> ERROR: One or more files did not pass the validity check!

danieltetraquark commented on 2019-09-08 13:58 (UTC)

@max.bra yes, indeed. (I should have checked that)

Lighttpd will listen to ipv4 only without this setting which can lead to problems.

max.bra commented on 2019-09-08 09:47 (UTC) (edited on 2019-09-08 09:49 (UTC) by max.bra)

hi danieltetraquark, we don't have any perl script in our lighttpd package.

Do you mean?

## Use IPv6?
server.use-ipv6 = "enable"

Anyway, this package installs an example lighttpd.conf and will never overwrite your official config. You can modify yours and it will be forever.

edit: markdown

danieltetraquark commented on 2019-09-05 16:50 (UTC)

Could you add the following to the lighttpd config to enable listening to ipv4 and ipv6? Like in the upstream/debian lighttpd config

# default listening port for IPv6 falls back to the IPv4 port
include_shell "/usr/share/lighttpd/ " + server.port

hakayova commented on 2019-07-19 00:03 (UTC) (edited on 2019-07-19 00:04 (UTC) by hakayova)

Thank you @max.bra. I followed your instructions and it worked. I was just relying on the AUR package for updates and assumed that when you wrote the patch to fix the issue, the package was also recreated with the patch. Maybe it was but it didn't work to re-install it via yaourt. Your method of compiling it with makepkg did. Thanks!

max.bra commented on 2019-07-18 06:22 (UTC) (edited on 2019-07-19 22:56 (UTC) by max.bra)

hi hakayova, i don't know what you mean with "I did not compile it as @max.bra recommended" but my recommendations are inside the PKGBUILD. the only way to make FTL works with nettle 3.5.x is to patch it. dot. download a snapshot (link in package AUR page), unzip and do a simple makepkg. after that i'm sure that everything will work. @all please do not rely on AUR helpers only, create your packages by hand too.

hakayova commented on 2019-07-18 04:54 (UTC) (edited on 2019-07-18 04:54 (UTC) by hakayova)

I cannot get pihole-FTL start with nettle 3.5.1-1. It runs as expected with the previous version (3.4.1-1). I did not compile it as @max.bra recommended but re-installed pi-hole-ftl from AUR package, thinking that the package may have been patched/updated, but it did not solve the problem and I had to re-downgrade the nettle to make it work again.

max.bra commented on 2019-07-11 06:34 (UTC)

hi katsuki, please do a full system update and recompile from snapshot pi-hole-ftl

Katsuki commented on 2019-07-11 03:26 (UTC)

I still got an error: "error while loading shared libraries: cannot open shared object file...." Pihole dns won't run

Treyarch commented on 2019-07-10 00:06 (UTC)

Looks like it now works and has been recompiled for the latest version.

graysky commented on 2019-07-07 13:38 (UTC) (edited on 2019-07-07 13:40 (UTC) by graysky)

Does recompiling against the new version restore function?

abumasood commented on 2019-07-07 13:14 (UTC)

nettle 3.5.1-1 update on 07/07/2019 broke pihole-FTL. Downgrade nettle to get pi-hole running again.

Treyarch commented on 2019-07-07 07:22 (UTC)

Bit of a warning, just did an update to libnettle and it stops pihole-FTL from starting.

Not sure if an actual fix I just downgraded my libnettle and it came back ¯_(ツ)_/¯ .

max.bra commented on 2019-07-06 15:29 (UTC)

Hi hakayova, indeed resolution and ping respond are separate problems. i'm able to resolve '' with both 'ping' and 'ping -4'. After that it respond to me. i don't know if there is a blacklist for ICMP attacks...

hakayova commented on 2019-07-06 00:39 (UTC) (edited on 2019-07-06 01:05 (UTC) by hakayova)

Dear all, I have been using pihole for almost 2 years now, however ping requests (,, etc.) are consistently failing 100% since the last update. Both pihole server and FTL server are running without reporting problems. Web browsing, package updating are not affected; however I seem to totally lost name resolution from CLI (ping works). Your guidance will be greatly appreciated. Edit: name resolution is probably not the issue here. The servers seem to have stopped responding to my ping requests. I am not sure why. It is now off-topic but can someone please confirm?

tdkl commented on 2019-06-30 14:45 (UTC) (edited on 2019-06-30 20:44 (UTC) by tdkl)

@graysky: I did use pikaur, now when doing manually it runs the prepare(), but fails later:

==> Creating package "pi-hole-server"... -> Generating .PKGINFO file... -> Generating .BUILDINFO file... -> Adding install file... -> Generating .MTREE file... -> Compressing package... /usr/share/makepkg/util/ line 36: 13287 Killed ${COMPRESSXZ[@]:-xz -c -z -} bsdtar: Write error ==> ERROR: Failed to create package file.

Maybe my 240MB RAM armv5 device is just too shitty.

graysky commented on 2019-06-30 09:31 (UTC)

Are you using an AUR helper? Builds fine with makepkg or makechrootpkg

tdkl commented on 2019-06-30 08:50 (UTC)

Getting this error with the most recent update:

==> Extracting sources... -> Extracting pihole-4.3.1.tar.gz with bsdtar -> Extracting admin-4.3.tar.gz with bsdtar ==> Starting prepare()... The next patch would create the file adlists.list, which already exists! Skipping patch. 1 out of 1 hunk ignored patching file advanced/Scripts/ patching file advanced/Scripts/ patching file advanced/Scripts/ patching file advanced/Scripts/ patching file advanced/Templates/logrotate patching file advanced/Templates/pihole.sudo patching file advanced/dnsmasq.conf.original The next patch would create the file dns-servers.conf, which already exists! Skipping patch. 1 out of 1 hunk ignored patching file patching file pihole ==> ERROR: A failure occurred in prepare(). Aborting...

tdkl commented on 2019-06-19 14:01 (UTC)

Managed to find the problem with the Long term data and Network not displaying. It needs pdo_sqlite and sqlite3 to be enabled in php.ini (somehow I didn't have both), now it displays the data.

max.bra commented on 2019-06-19 11:31 (UTC)

Hi PS0, it's FTL pkg that conflict with dnsmasq

PS0 commented on 2019-06-19 11:28 (UTC)

This package conflicts with dnsmasq (not mentioned in PKGBUILD).

tdkl commented on 2019-06-18 21:58 (UTC) (edited on 2019-06-18 22:01 (UTC) by tdkl)

I already have it disabled ...

I noticed it throws the same error when going to Long term data / Query log (/db_queries.php), but Query log in the top menu displays fine.

I tried resetting the pihole-ftl.db in /etc/pihole, but that didn't do anything.

graysky commented on 2019-06-18 21:02 (UTC)

I think I solved this by disabling systemd-resolved and just manually setting /etc/resolv.conf with the correct nameserver... I seem to recall seeing something in the journal about port 53 already being bound. Been too long and covered too many miles since.

tdkl commented on 2019-06-18 19:17 (UTC) (edited on 2019-06-18 19:24 (UTC) by tdkl)

@graysky: I also have the same behaviour as you, although I get an error "An unknown error occured while loading the data". Running on archv5 pogoplug v2.

Pi-hole is the only DHCP server on my network, but also have an odd issue when saving a Static DHCP Lease in DHCP :

Info A new static address has been added

Error There was a problem applying your settings. Debugging information: PHP error (2): fopen(/etc/dnsmasq.d/04-pihole-static-dhcp.conf): failed to open stream: Operation not permitted in /srv/http/pihole/admin/scripts/pi-hole/php/savesettings.php:66

The file has 644 root:root permissions. The change is actually saved, the entry is never displayed on the Static lease list.

Maybe both of them are connected, that some info can't be read to be displayed on the website? Otherwise the DNS queries are logged and displayed.

[EDIT] I've managed to solve the DHCP Static lease error, the open_basedir wasn't set correctly for the whole dnsmasq.d folder. The no data on network page remains.

max.bra commented on 2019-05-19 17:26 (UTC)

@graysky: anyway, i use pi-hole as DHCP server...

max.bra commented on 2019-05-19 17:22 (UTC)

@graysky: yes "network overview" is populated for me.

EDIT: is that only available if using pihole as a DHCP server?

uhmm, no. seems to be based on queries:

graysky commented on 2019-05-19 10:06 (UTC) (edited on 2019-05-19 10:08 (UTC) by graysky)

@max.bra - Is the "network overview" section getting populated for you? I have no data in that section. Everything else seems in functional.

EDIT: is that only available if using pihole as a DHCP server?

max.bra commented on 2019-04-28 07:47 (UTC) (edited on 2019-04-28 07:51 (UTC) by max.bra)

@graysky my last secret agency security report was reporting you abandon nginx in favour of lighttpd...

anyway, reverting is not a problem (edit) ... reverted, waiting for mreutman to better investigate the problem. sorry for the noise.

graysky commented on 2019-04-28 07:41 (UTC)

@max.bra - I don't know if that change is needed... I am running nginx without issue with fastcgi_pass; configured per the wiki on an ODROID-C2 (Arch ARM).

For example, I do not have /var/run/php-fpm/php-fpm.sock on my system running this package with nginx at all, just an empty directory.

Would you revert the change and comment on the wiki? I do not know the scope of this change for other users...

@mreutman - suspect you overlooked something/should not need that modification.

max.bra commented on 2019-04-28 07:13 (UTC)

@mreutman thanks a lot for reporting!! nginx is, at the moment, a little bit under tested. updated w/o version bump.

mreutman commented on 2019-04-25 00:31 (UTC)

I believe I have run into an error when getting this running on a Raspberry Pi 3 B+ using Nginx. I followed the instructions on the Wiki as far as I know; if you see something I overlooked, please let me know.

In particular, it looks like the php fastcgi_pass field in the conf file is not set correctly to allow the web interface to work out of the box. I had to change it from "fastcgi_pass;" to "fastcgi_pass unix:/var/run/php-fpm/php-fpm.sock;" in order for things to work. Without that change, Nginx would serve up "502 Bad Gateway".

Not sure if this is something that can be changed in the package or if it should just be documented in the Wiki page.

max.bra commented on 2019-04-07 19:59 (UTC)

hi wshuman3, thanks for reporting but psmisc it's from base group and it's assumed default installed. same for base-devel group.

wshuman3 commented on 2019-04-07 16:29 (UTC) (edited on 2019-04-07 16:30 (UTC) by wshuman3)

This should include psmisc as a dependency, /usr/bin/pihole references killall which is included in psmisc.

max.bra commented on 2019-02-19 20:33 (UTC)

I was just wondering if the problem was only on ftl... :-D

jshap commented on 2019-02-19 20:27 (UTC) (edited on 2019-02-19 20:27 (UTC) by jshap)

same issues as in pi-hole-ftl, the links in the PKGBUILD need instead to be

max.bra commented on 2019-02-12 16:08 (UTC)

yes it is! "our" 01-pihole.conf is without dns server and populated during first install. upstream do the same with their install script.

i never replace conf files with .pacnew ones without a check. it's better to "human" merge them with diff like applications.

glad you solved and thanks for reporting.

mystique commented on 2019-02-12 16:06 (UTC)


dnsmasq doesn't do encrypted dns (of any type) only straight udp/tcp port 53.

You would need to use something else (dnscrypt-proxy) for the service of encrypted dns.

I would also recommend quad9 vs cloudflare, but you should do you own homework for your own reasons..

And this would be a forum thing (imho) rather than a package discussion..

My 0.02

erkexzcx commented on 2019-02-12 16:00 (UTC)

So basically what I've done was "mv /etc/dnsmasq.d/01-pihole.conf.pacnew /etc/dnsmasq.d/01-pihole.conf" and rebooted (I think). It stopped working.

Now I just executed "pihole -a setdns" and settings were overridden in pi-hole. Everything started to work fine.

But because I was using cloudflared DOH DNS service on my computer - I might guess it is the root cause, but it is not - I went to pi-hole settings and put cloudflared settings (they are "") and once again everything works.

So probably the root cause was messing with "/etc/dnsmasq.d/01-pihole.conf.pacnew"... :)

erkexzcx commented on 2019-02-12 11:55 (UTC)

Sure thing. Will check!

max.bra commented on 2019-02-12 08:44 (UTC) (edited on 2019-02-12 08:44 (UTC) by max.bra)

@erkexzcx oh, about wiki changes: as stated above the server package is not officially upstream supported as the standalone version is.

max.bra commented on 2019-02-12 08:42 (UTC)

@erkexzcx rev. 3 changes cannot lead to DNS resolution loss. i recommend you to check your .pacnew files and this (

erkexzcx commented on 2019-02-12 07:17 (UTC)

Looks like pi-hole-server 4.2.1-3 no longer works as DNS server. pi-hole-server 4.2.1-2 worked fine. Now my DNS queries are not being resolved by Pi-Hole.

max.bra commented on 2019-02-11 11:55 (UTC)

@Taijian yes, it was a forgetfulness!! thanks for reporting.

Taijian commented on 2019-02-11 11:26 (UTC)

Looking at this package again and checking against the 'official' changelog for 4.2, I noticed that the fix for the wpad vulnerability included in v4.2+ is not included in this package because you opted to not ship the 01-pihole.conf file included with the upstream package but instead replace it with your homebrew version 'dnsmasq.include', which hasn't been updated with this fix. Is there a reason for this that I'm just not seeing, or is this an oversight?

jonandermb commented on 2019-02-08 19:05 (UTC)

Crashing for me here: The -2 patch seemed to mitigate the symptoms a bit, but I get pihole rebooted every 5 minutes or so :(

DriverChief commented on 2019-02-05 19:02 (UTC)

It seems to be working for me as well. I've seen a few resizes occur without a crash after running for a few hours. Thank you!

mystique commented on 2019-02-05 16:55 (UTC) (edited on 2019-02-05 17:00 (UTC) by mystique)

this is a different machine, same env..

(formatted) (/formatted)

[2019-02-05 11:43:40.691] Resizing "/FTL-queries" from 79036416 to 79331328 [2019-02-05 11:43:40.737] Resizing "/FTL-strings" from 4382720 to 4386816 [2019-02-05 11:43:40.883] Resizing "/FTL-strings" from 4386816 to 4390912 [2019-02-05 11:43:40.918] Resizing "/FTL-queries" from 79331328 to 79626240 [2019-02-05 11:43:41.037] Resizing "/FTL-strings" from 4390912 to 4395008 [2019-02-05 11:43:41.061] Imported 1104617 queries from the long-term database [2019-02-05 11:43:41.061] -> Total DNS queries: 1104617 [2019-02-05 11:43:41.061] -> Cached DNS queries: 347654 [2019-02-05 11:43:41.061] -> Forwarded DNS queries: 333982 [2019-02-05 11:43:41.061] -> Exactly blocked DNS queries: 422981 [2019-02-05 11:43:41.061] -> Unknown DNS queries: 0 [2019-02-05 11:43:41.061] -> Unique domains: 72655 [2019-02-05 11:43:41.062] -> Unique clients: 3 [2019-02-05 11:43:41.062] -> Known forward destinations: 1 [2019-02-05 11:43:41.062] Successfully accessed setupVars.conf [2019-02-05 11:43:41.092] PID of FTL process: 374 [2019-02-05 11:43:41.092] Listening on port 4711 for incoming IPv4 telnet connections [2019-02-05 11:43:41.092] Listening on Unix socket [2019-02-05 11:43:41.101] Compiled 0 Regex filters and 14 whitelisted domains in 0.1 msec (0 errors) [2019-02-05 11:43:41.105] /etc/pihole/black.list: parsed 0 domains (took 0.0 ms) [2019-02-05 11:43:41.488] /etc/pihole/gravity.list: parsed 126293 domains (took 377.2 ms) [2019-02-05 11:43:49.749] Resizing "/FTL-queries" from 79626240 to 79921152 [2019-02-05 11:43:57.435] Resizing "/FTL-strings" from 4395008 to 4399104 [2019-02-05 11:44:58.785] Resizing "/FTL-strings" from 4399104 to 4403200 [2019-02-05 11:45:10.911] Resizing "/FTL-queries" from 79921152 to 80216064 [2019-02-05 11:46:07.925] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [2019-02-05 11:46:07.925] ----------------------------> FTL crashed! <---------------------------- [2019-02-05 11:46:07.925] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [2019-02-05 11:46:07.925] Please report a bug at [2019-02-05 11:46:07.925] and include in your report already the following details: [2019-02-05 11:46:07.925] FTL has been running for 183 seconds [2019-02-05 11:46:07.925] FTL branch: master [2019-02-05 11:46:07.925] FTL version: v4.2 [2019-02-05 11:46:07.925] FTL commit: aca02c4 [2019-02-05 11:46:07.925] FTL date: 2019-02-03 [2019-02-05 11:46:07.925] FTL user: root [2019-02-05 11:46:07.925] Received signal: Segmentation fault [2019-02-05 11:46:07.925] at address: 140136641380264 [2019-02-05 11:46:07.925] with code: SEGV_MAPERR (Address not mapped to object) [2019-02-05 11:46:07.926] Backtrace: [2019-02-05 11:46:07.926] B[0000]: /usr/bin/pihole-FTL(+0x17d37) [0x556704054d37] [2019-02-05 11:46:07.926] B[0001]: /usr/lib/ [0x7f74200403c0] [2019-02-05 11:46:07.926] B[0002]: /usr/bin/pihole-FTL(FTL_new_query+0x34b) [0x55670405e1fb] [2019-02-05 11:46:07.926] B[0003]: /usr/bin/pihole-FTL(tcp_request+0x65a) [0x55670406f61a] [2019-02-05 11:46:07.926] B[0004]: /usr/bin/pihole-FTL(+0x45816) [0x556704082816] [2019-02-05 11:46:07.926] B[0005]: /usr/bin/pihole-FTL(main_dnsmasq+0x1188) [0x556704084418] [2019-02-05 11:46:07.926] B[0006]: /usr/bin/pihole-FTL(main+0xce) [0x55670405326e] [2019-02-05 11:46:07.926] B[0007]: /usr/lib/ [0x7f741fe8e223] [2019-02-05 11:46:07.926] B[0008]: /usr/bin/pihole-FTL(_start+0x2e) [0x5567040533ae] [2019-02-05 11:46:07.926] Thank you for helping us to improve our FTL engine! [2019-02-05 11:46:07.926] FTL terminated! [2019-02-05 11:46:26.876] Resizing "/FTL-strings" from 4403200 to 4407296 [2019-02-05 11:46:49.395] Resizing "/FTL-queries" from 80216064 to 80510976

esxi 6.x vm, dual core, 2GB ram.

[2019-02-05 11:43:40.691] Resizing "/FTL-queries" from 79036416 to 79331328 [2019-02-05 11:43:40.737] Resizing "/FTL-strings" from 4382720 to 4386816 [2019-02-05 11:43:40.883] Resizing "/FTL-strings" from 4386816 to 4390912 [2019-02-05 11:43:40.918] Resizing "/FTL-queries" from 79331328 to 79626240 [2019-02-05 11:43:41.037] Resizing "/FTL-strings" from 4390912 to 4395008 [2019-02-05 11:43:41.061] Imported 1104617 queries from the long-term database [2019-02-05 11:43:41.061] -> Total DNS queries: 1104617 [2019-02-05 11:43:41.061] -> Cached DNS queries: 347654 [2019-02-05 11:43:41.061] -> Forwarded DNS queries: 333982 [2019-02-05 11:43:41.061] -> Exactly blocked DNS queries: 422981 [2019-02-05 11:43:41.061] -> Unknown DNS queries: 0 [2019-02-05 11:43:41.061] -> Unique domains: 72655 [2019-02-05 11:43:41.062] -> Unique clients: 3 [2019-02-05 11:43:41.062] -> Known forward destinations: 1 [2019-02-05 11:43:41.062] Successfully accessed setupVars.conf [2019-02-05 11:43:41.092] PID of FTL process: 374 [2019-02-05 11:43:41.092] Listening on port 4711 for incoming IPv4 telnet connections [2019-02-05 11:43:41.092] Listening on Unix socket [2019-02-05 11:43:41.101] Compiled 0 Regex filters and 14 whitelisted domains in 0.1 msec (0 errors) [2019-02-05 11:43:41.105] /etc/pihole/black.list: parsed 0 domains (took 0.0 ms) [2019-02-05 11:43:41.488] /etc/pihole/gravity.list: parsed 126293 domains (took 377.2 ms) [2019-02-05 11:43:49.749] Resizing "/FTL-queries" from 79626240 to 79921152 [2019-02-05 11:43:57.435] Resizing "/FTL-strings" from 4395008 to 4399104 [2019-02-05 11:44:58.785] Resizing "/FTL-strings" from 4399104 to 4403200 [2019-02-05 11:45:10.911] Resizing "/FTL-queries" from 79921152 to 80216064 [2019-02-05 11:46:07.925] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [2019-02-05 11:46:07.925] ----------------------------> FTL crashed! <---------------------------- [2019-02-05 11:46:07.925] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [2019-02-05 11:46:07.925] Please report a bug at [2019-02-05 11:46:07.925] and include in your report already the following details: [2019-02-05 11:46:07.925] FTL has been running for 183 seconds [2019-02-05 11:46:07.925] FTL branch: master [2019-02-05 11:46:07.925] FTL version: v4.2 [2019-02-05 11:46:07.925] FTL commit: aca02c4 [2019-02-05 11:46:07.925] FTL date: 2019-02-03 [2019-02-05 11:46:07.925] FTL user: root [2019-02-05 11:46:07.925] Received signal: Segmentation fault [2019-02-05 11:46:07.925] at address: 140136641380264 [2019-02-05 11:46:07.925] with code: SEGV_MAPERR (Address not mapped to object) [2019-02-05 11:46:07.926] Backtrace: [2019-02-05 11:46:07.926] B[0000]: /usr/bin/pihole-FTL(+0x17d37) [0x556704054d37] [2019-02-05 11:46:07.926] B[0001]: /usr/lib/ [0x7f74200403c0] [2019-02-05 11:46:07.926] B[0002]: /usr/bin/pihole-FTL(FTL_new_query+0x34b) [0x55670405e1fb] [2019-02-05 11:46:07.926] B[0003]: /usr/bin/pihole-FTL(tcp_request+0x65a) [0x55670406f61a] [2019-02-05 11:46:07.926] B[0004]: /usr/bin/pihole-FTL(+0x45816) [0x556704082816] [2019-02-05 11:46:07.926] B[0005]: /usr/bin/pihole-FTL(main_dnsmasq+0x1188) [0x556704084418] [2019-02-05 11:46:07.926] B[0006]: /usr/bin/pihole-FTL(main+0xce) [0x55670405326e] [2019-02-05 11:46:07.926] B[0007]: /usr/lib/ [0x7f741fe8e223] [2019-02-05 11:46:07.926] B[0008]: /usr/bin/pihole-FTL(_start+0x2e) [0x5567040533ae] [2019-02-05 11:46:07.926] Thank you for helping us to improve our FTL engine! [2019-02-05 11:46:07.926] FTL terminated! [2019-02-05 11:46:26.876] Resizing "/FTL-strings" from 4403200 to 4407296 [2019-02-05 11:46:49.395] Resizing "/FTL-queries" from 80216064 to 80510976

It did crash, but it spawned another and kept going..

dnsadmn@arch98c1e8 /etc> cat -n dnsmasq.d/* | egrep -v "#" 22
23 addn-hosts=/etc/pihole/gravity.list 24 addn-hosts=/etc/pihole/local.list 25 addn-hosts=/etc/pihole/black.list 26 localise-queries 27 no-resolv 28 cache-size=10000 29 log-queries=extra 30 log-facility=/run/log/pihole/pihole.log 31 local-ttl=2 32 log-async 33 server= 34 server= 35 interface=ens192 36
37 all-servers 38 edns-packet-max=4096

23-35 is pihole config and 37-38 is a local.conf that I added to get my 33-34 recursors going. The edns is for the squid proxies directly querying this machine.

mystique commented on 2019-02-05 15:02 (UTC) (edited on 2019-02-05 15:22 (UTC) by mystique)

[2019-02-05 10:01:14.786] ########## FTL started! ##########
[2019-02-05 10:01:14.786] FTL branch: master
[2019-02-05 10:01:14.786] FTL version: v4.2
[2019-02-05 10:01:14.787] FTL commit: aca02c4
[2019-02-05 10:01:14.787] FTL date: 2019-02-03
[2019-02-05 10:01:14.787] FTL user: root

Will report back if there are issues.


graysky commented on 2019-02-05 14:58 (UTC)

@RyanFlaherty @mystique - Please try pi-hole-ftl 4.2-2

graysky commented on 2019-02-05 14:57 (UTC)

@max.bra - Seems to be fixed for me. I will push to the AUR. BTW, you can trigger the bug by browsing to sites with a lot of ads at least on ARM hardware.

For example:

Here is my log tailed with the patches:

% tail -f /run/log/pihole-ftl/pihole-FTL.log 
% tail -f /run/log/pihole-ftl/pihole-FTL.log 
[2019-02-05 09:53:47.023]  -> Unique clients: 8
[2019-02-05 09:53:47.023]  -> Known forward destinations: 1
[2019-02-05 09:53:47.023] Successfully accessed setupVars.conf
[2019-02-05 09:53:47.027] PID of FTL process: 141
[2019-02-05 09:53:47.027] Listening on port 4711 for incoming IPv4 telnet connections
[2019-02-05 09:53:47.027] Listening on port 4711 for incoming IPv6 telnet connections
[2019-02-05 09:53:47.027] Listening on Unix socket
[2019-02-05 09:53:47.029] Compiled 0 Regex filters and 29 whitelisted domains in 0.1 msec (0 errors)
[2019-02-05 09:53:47.029] /etc/pihole/black.list: parsed 4 domains (took 0.1 ms)
[2019-02-05 09:53:49.223] /etc/pihole/gravity.list: parsed 264453 domains (took 2193.6 ms)
[2019-02-05 09:55:44.196] Resizing "/FTL-strings" from 24576 to 28672
[2019-02-05 09:55:51.190] Resizing "/FTL-strings" from 28672 to 32768

mystique commented on 2019-02-05 14:50 (UTC) (edited on 2019-02-05 15:23 (UTC) by mystique)

egrep -v "^#" /etc/security/limits.conf

proxy soft nofile 65536
proxy hard nofile 65536

* soft nofile 65536
* hard nofile 65536

(don't hate please.. runit-systemd) I needed it to keep running regardless of the crash until I could work something out..

cat /var/service/pihole-ftl/run

#!/bin/sh -e

exec chpst -o 65536 chpst -e env \
/usr/bin/pihole-FTL -f

and in env I just set TZ=/etc/localtime


graysky commented on 2019-02-05 14:48 (UTC)

@max.bra - I am testing the two commits @mystique referenced now on my ARM device. No crash yet. If it's good, I will push to pi-hole-ftl-4.2-2 which will incorporate them. It seems as though upstream has a 4.2.1 milestone queued up with these fixes:

max.bra commented on 2019-02-05 14:47 (UTC)

for me default limits.conf, systemd/system.conf and systemd/user.conf...

max.bra commented on 2019-02-05 14:44 (UTC)

still no crash so far!!! seems to be some user limit... any tips?

graysky commented on 2019-02-05 14:39 (UTC)

@max.bra - I am experiencing this bug as well. Link to /run/log/pihole-ftl/pihole-FTL.log:

mystique commented on 2019-02-05 14:36 (UTC)

so far a fix for me has been this:

Add these lines to a new config file in /etc/dnsmasq.d/:


And restart FTL: sudo service pihole-FTL restart

And no crash that I can see.

[2019-02-05 09:02:33.862] Resizing "/FTL-strings" from 8192 to 12288 [2019-02-05 09:10:55.157] Resizing "/FTL-strings" from 12288 to 16384 [2019-02-05 09:22:47.979] Resizing "/FTL-strings" from 16384 to 20480

max.bra commented on 2019-02-05 10:03 (UTC)

Hi mystique, for an FTL problem you should post to FTL package. Anyway, it's a problem, i have same hardware and not a crash for now...

DriverChief commented on 2019-02-04 23:54 (UTC) (edited on 2019-02-05 00:12 (UTC) by DriverChief)

@graysky Forgive me if I don't provide you what you need right away. I'm not a frequent poster on AUR and not really a wiz enough to maintain my own packages.

Update: looks like this is a known issue upstream

I see this behavior on my pi3b+ running arch linux on arm. my logs have it regularly failing on realloc_shm() after varying delays on startup. I notice pihole starts as root, but then when checking the webinterface its running as nobody. Could it be that when the user gets changed to 'nobody' it no longer can open and resize the shared memory object it needs?

edit: adding log excerpt below

[2019-02-04 18:04:57.170] Using log file /run/log/pihole-ftl/pihole-FTL.log
[2019-02-04 18:04:57.170] ########## FTL started! ##########
[2019-02-04 18:04:57.170] FTL branch: master
[2019-02-04 18:04:57.170] FTL version: v4.2
[2019-02-04 18:04:57.171] FTL commit: a636945
[2019-02-04 18:04:57.171] FTL date: 2019-02-03
[2019-02-04 18:04:57.171] FTL user: root
[2019-02-04 18:04:57.171] WARNING: Starting pihole-FTL as user root is not recommended
[2019-02-04 18:04:57.171] Starting config file parsing (/etc/pihole/pihole-FTL.conf)
[2019-02-04 18:04:57.171] SOCKET_LISTENING: only local
[2019-02-04 18:04:57.171] AAAA_QUERY_ANALYSIS: Show AAAA queries
[2019-02-04 18:04:57.171] MAXDBDAYS: max age for stored queries is 365 days
[2019-02-04 18:04:57.171] RESOLVE_IPV6: Resolve IPv6 addresses
[2019-02-04 18:04:57.171] RESOLVE_IPV4: Resolve IPv4 addresses
[2019-02-04 18:04:57.171] DBINTERVAL: saving to DB file every minute
[2019-02-04 18:04:57.171] DBFILE: Using /etc/pihole/pihole-FTL.db
[2019-02-04 18:04:57.172] MAXLOGAGE: Importing up to 24.0 hours of log data
[2019-02-04 18:04:57.172] PRIVACYLEVEL: Set to 0
[2019-02-04 18:04:57.172] IGNORE_LOCALHOST: Show queries from localhost
[2019-02-04 18:04:57.172] BLOCKINGMODE: Null IPs for blocked domains
[2019-02-04 18:04:57.172] REGEX_DEBUGMODE: Inactive
[2019-02-04 18:04:57.172] ANALYZE_ONLY_A_AND_AAAA: Disabled. Analyzing all queries
[2019-02-04 18:04:57.172] DBIMPORT: Importing history from database
[2019-02-04 18:04:57.172] PIDFILE: Using /run/pihole-ftl/
[2019-02-04 18:04:57.172] PORTFILE: Using /run/pihole-ftl/pihole-FTL.port
[2019-02-04 18:04:57.172] SOCKETFILE: Using /run/pihole-ftl/FTL.sock
[2019-02-04 18:04:57.172] WHITELISTFILE: Using /etc/pihole/whitelist.txt
[2019-02-04 18:04:57.172] BLACKLISTFILE: Using /etc/pihole/black.list
[2019-02-04 18:04:57.172] GRAVITYFILE: Using /etc/pihole/gravity.list
[2019-02-04 18:04:57.173] REGEXLISTFILE: Using /etc/pihole/regex.list
[2019-02-04 18:04:57.173] SETUPVARSFILE: Using /etc/pihole/setupVars.conf
[2019-02-04 18:04:57.173] AUDITLISTFILE: Using /etc/pihole/auditlog.list
[2019-02-04 18:04:57.173] Finished config file parsing
[2019-02-04 18:04:57.173] Compiled 0 Regex filters and 0 whitelisted domains in 0.1 msec (0 errors)
[2019-02-04 18:04:57.174] Database successfully initialized
[2019-02-04 18:04:57.175] New forward server: (0/4096)
[2019-02-04 18:04:57.176] New forward server: (1/4096)
[2019-02-04 18:04:57.181] Imported 1024 queries from the long-term database
[2019-02-04 18:04:57.181] -> Total DNS queries: 1024
[2019-02-04 18:04:57.181] -> Cached DNS queries: 205
[2019-02-04 18:04:57.181] -> Forwarded DNS queries: 487
[2019-02-04 18:04:57.182] -> Exactly blocked DNS queries: 180
[2019-02-04 18:04:57.182] -> Unknown DNS queries: 152
[2019-02-04 18:04:57.182] -> Unique domains: 161
[2019-02-04 18:04:57.182] -> Unique clients: 3
[2019-02-04 18:04:57.182] -> Known forward destinations: 2
[2019-02-04 18:04:57.182] Successfully accessed setupVars.conf
[2019-02-04 18:04:57.188] PID of FTL process: 13980
[2019-02-04 18:04:57.188] Listening on port 4711 for incoming IPv4 telnet connections
[2019-02-04 18:04:57.188] Listening on port 4711 for incoming IPv6 telnet connections
[2019-02-04 18:04:57.189] Listening on Unix socket
[2019-02-04 18:04:57.190] Compiled 0 Regex filters and 0 whitelisted domains in 0.1 msec (0 errors)
[2019-02-04 18:04:57.190] /etc/pihole/black.list: parsed 0 domains (took 0.0 ms)
[2019-02-04 18:04:58.414] /etc/pihole/gravity.list: parsed 112784 domains (took 1223.4 ms)
[2019-02-04 18:04:58.414] Resizing "/FTL-strings" from 4096 to 8192
[2019-02-04 18:04:58.414] realloc_shm(): Failed to open shared memory object "/FTL-strings": Permission denied
[2019-02-04 18:04:58.415] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
[2019-02-04 18:04:58.415] ----------------------------> FTL crashed! <----------------------------
[2019-02-04 18:04:58.415] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
[2019-02-04 18:04:58.415] Please report a bug at
[2019-02-04 18:04:58.415] and include in your report already the following details:

[2019-02-04 18:04:58.415] FTL has been running for 1 seconds
[2019-02-04 18:04:58.415] FTL branch: master
[2019-02-04 18:04:58.415] FTL version: v4.2
[2019-02-04 18:04:58.415] FTL commit: a636945
[2019-02-04 18:04:58.415] FTL date: 2019-02-03
[2019-02-04 18:04:58.415] FTL user: root
[2019-02-04 18:04:58.415] Received signal: Segmentation fault
[2019-02-04 18:04:58.415] at address: 1995751425
[2019-02-04 18:04:58.415] with code: SEGV_MAPERR (Address not mapped to object)
[2019-02-04 18:04:58.415] Backtrace:
[2019-02-04 18:04:58.415] B[0000]: /usr/bin/pihole-FTL(+0xf6dc) [0x48f6dc]
[2019-02-04 18:04:58.415] B[0001]: /usr/lib/ [0x76c9fe10]
[2019-02-04 18:04:58.415] Thank you for helping us to improve our FTL engine!
[2019-02-04 18:04:58.415] FTL terminated!

graysky commented on 2019-02-04 23:29 (UTC)

@mystique - You provided no useful info beyond some log output. What webserver? How did you trigger the error?

mystique commented on 2019-02-04 23:06 (UTC) (edited on 2019-02-04 23:08 (UTC) by mystique)

dnsadmn@vmArch59 ~> pacman -Q | grep pi-hole pi-hole-ftl 4.2-1 pi-hole-server 4.2.1-1

dnsadmn@vmArch59 ~> pihole version Pi-hole version is 4.1 (Latest: v4.2.1) FTL version is v4.2 (Latest: v4.2)

sh -x /usr/bin/pihole + /opt/pihole/ Pi-hole version is 4.1 (Latest: v4.2.1) FTL version is v4.2 (Latest: v4.2)

[2019-02-04 18:04:10.653] create_shm(): shm_unlink("/FTL-client-0") failed: Permission denied (13) [2019-02-04 18:04:10.653] create_shm(): Failed to create_shm shared memory object "/FTL-client-0": File exists [2019-02-04 18:04:10.653] Failed to initialize new overTime client 0 [2019-02-04 18:04:10.653] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [2019-02-04 18:04:10.653] ----------------------------> FTL crashed! <---------------------------- [2019-02-04 18:04:10.653] !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [2019-02-04 18:04:10.653] Please report a bug at [2019-02-04 18:04:10.653] and include in your report already the following details:

[2019-02-04 18:04:10.653] FTL has been running for 10 seconds [2019-02-04 18:04:10.653] FTL branch: master [2019-02-04 18:04:10.653] FTL version: v4.2 [2019-02-04 18:04:10.653] FTL commit: a636945 [2019-02-04 18:04:10.653] FTL date: 2019-02-03 [2019-02-04 18:04:10.653] FTL user: root [2019-02-04 18:04:10.653] Received signal: Segmentation fault [2019-02-04 18:04:10.653] at address: 0 [2019-02-04 18:04:10.653] with code: SEGV_MAPERR (Address not mapped to object) [2019-02-04 18:04:10.654] Backtrace: [2019-02-04 18:04:10.654] B[0000]: /usr/bin/pihole-FTL(+0x17d37) [0x5583d8771d37]

Completely stock build out of the box.. they changed things in their 4.2 release..

I'm still looking for the fix.. but wanted to share this this seems to not work..

max.bra commented on 2019-02-02 17:08 (UTC)

@ghost33 no new version or at least not yet officially released

max.bra commented on 2019-01-21 08:54 (UTC)

@pepper_chico that is not for pihole-server, it's for FTL

pepper_chico commented on 2019-01-21 08:51 (UTC)

On ArchLinuxARM, I had to apply the following manually for pihole to start:

DNSStubListener=no on resolved.conf

ghostv33 commented on 2018-12-30 22:32 (UTC)

Got it to work Forgot to copy the config file of lighttpd as stated in the wiki page And thanks for bringing pihole to arch :D Debian is slow so gave up then found this page ❤

max.bra commented on 2018-12-30 22:25 (UTC) has nothing to do with us. i don't know where that url come from but is not from this package config files for sure.

http://pi.hole/admin/ or http://<pihole_machine_ip>/admin/ or http://<localhost_127.0.0.1>/admin/ should work out of the box.

i think you'd better follow pi-hole ARCH wiki:

ghostv33 commented on 2018-12-30 21:40 (UTC)

Ya I did Webpage shows 404 Not Found

max.bra commented on 2018-12-28 22:30 (UTC)

hi ghostv33, pihole web admin is present. stupid question: did you start lighttpd service?

ghostv33 commented on 2018-12-28 15:31 (UTC)

I cant seem to access the web panel in the browser?is it not available in the arch one?

max.bra commented on 2018-12-12 08:13 (UTC)

yes thanks, was working on it. anyway, we know how to do a checksum, a report is enough.

aphirst commented on 2018-12-12 08:06 (UTC) (edited on 2018-12-12 08:06 (UTC) by aphirst)

1a6aa653fdf7a6406d550aabb023cafe arch-server-admin-4.1.patch

172047cda180834a3f9f8325359e65e3 arch-server-core-4.1.patch

It seems those 2 patches updated again.

max.bra commented on 2018-12-10 15:56 (UTC)

@graysky had time to update only today! all is functional here. thanks for the update.

graysky commented on 2018-12-10 01:11 (UTC)

4.1-2 is fine... nothing to change.

LordSandwurm commented on 2018-12-09 23:35 (UTC)

Hi The package cant be build. The md5 sums ar wrong in the PKGBUILD file. Afte changing them the package builds fine.

graysky commented on 2018-12-09 22:26 (UTC)

@max.bra - I updated both packages for 4.1 just now. This package just needs to have /usr/bin/dig so I added bind-tools to the deps. Please review and comment.

max.bra commented on 2018-11-28 20:47 (UTC)

markdown parenthesis escaping ---> hardest work!!! :-D

graysky commented on 2018-11-28 20:44 (UTC)

and a localized good one!

ottimo lavoro!

max.bra commented on 2018-11-28 20:38 (UTC) (edited on 2018-11-28 20:44 (UTC) by max.bra)

@graysky :D and a localized good one!

graysky commented on 2018-11-28 20:22 (UTC) - Good citizen!

max.bra commented on 2018-11-28 09:28 (UTC)

@graysky public service accomplished :D

graysky commented on 2018-11-28 09:17 (UTC)

@max.bra - Think of it as a public service :D

max.bra commented on 2018-11-28 09:09 (UTC)

@graysky i will but this is a generic linux configuration issue that has nothing to do with pi-hole.

graysky commented on 2018-11-28 09:05 (UTC) (edited on 2018-11-28 09:06 (UTC) by graysky)

@max.bra - You might want to comment on that github issue. I found it by googling the error so it's reasonable to think that others will too.

Also, filesystem-2018.8-1 ships with an empty /etc/hosts

max.bra commented on 2018-11-28 08:00 (UTC) (edited on 2018-11-28 08:05 (UTC) by max.bra)

yes but that comments are not correct.

$ cat /etc/resolv.conf
# Generated by resolvconf
search domain.local

assuming 'archpi' as the local hostname

$ cat /etc/hosts           localhost.domain.local   localhost
::1                 localhost.domain.local   localhost

<lanip_v4>          archpi.domain.local     archpi
<lanip_v6>          archpi.domain.local     archpi is localhost and should never be the hostname

edit: corrected localhost domain

graysky commented on 2018-11-27 20:06 (UTC) (edited on 2018-11-27 22:22 (UTC) by graysky)

This is affecting out package:

% pihole -g
  [✗] DNS resolution is currently unavailable
  [i] Time until retry: 85

The solution is to add an entry to /etc/hosts BUT doing so breaks the ability of the webui to resolve local IP addresses to local hostnames, at least with my setup.


max.bra commented on 2018-10-29 08:17 (UTC) (edited on 2018-10-29 08:18 (UTC) by max.bra)

hi lmhp82, do you have some free ram memory and/or some swap? it seems you have reached some kind of limit in your box.

free -m
ulimit -a

will help you to find out your problem.

lmhp82 commented on 2018-10-29 00:59 (UTC)

I’m running pi-hole on a Raspberry Pi Zero W on Arch Linux. Everything seems to be running correctly except when running pihole -g

Expected Behaviour: Update ad serving domains

Actual Behaviour: $ pihole -g [i] Neutrino emissions detected… /opt/pihole/ xmalloc: cannot allocate 93249269 bytes /opt/pihole/ xrealloc: cannot allocate 94812160 bytes $

Debug Token: Pi-hole debug options are disabled.

ArchLinux distribution of Pi-hole is not supported by official project and this debug report CAN NOT be uploaded to their debug team. Follow instructions carefully on the project wiki page ( 1) and please refer to AUR package web page for support and debugging ( 1).

max.bra commented on 2018-10-15 15:52 (UTC)

no, FTL should not remain stopped. pihole script "unlink" blocklists and restart FTL (dnsmasq) to block zero IP, wait for requested timeout then "relink" blocklists and restart FTL again to reload blocked IPs.

my web admin page show 0 domain blocked, the countdown and meantime all is usable.

any syslog relevant? are you in a container?

graysky commented on 2018-10-15 14:34 (UTC) (edited on 2018-10-15 14:36 (UTC) by graysky)

@max.bra - If I hit the 'disable for 5 minutes' the entire web GUI freezes up for about 25 seconds. When it unfreezes, I show 0 domains on the block list and FTL is offline (the offline bit is expected, no)? When 5 min is up, FTL reloads and the normal amount of blocklisted domains returns.

max.bra commented on 2018-10-15 13:52 (UTC)

@graysky seems to work here. your behavior?

graysky commented on 2018-10-15 12:11 (UTC)

@max.bra - Any thoughts on how to get the "disable pihole for xxx min" to work under 4.x?

max.bra commented on 2018-09-18 07:24 (UTC)

@tsago updated without version bump. thanks for reporting.

tsago commented on 2018-09-17 20:38 (UTC)

Hi there! On my raspi, the default lighttpd config for pi-hole fails to start (just re-installed from scratch), systemctl logs:
Sep 17 21:53:13 alarmpi lighttpd-angel[270]: 2018-09-17 21:53:13: (configfile.c.1451) pclose "cat external.conf 2>/dev/null"failed: Success
Sep 17 21:53:13 alarmpi lighttpd-angel[270]: 2018-09-17 21:53:13: (configfile.c.1289) source: /etc/lighttpd/lighttpd.conf line: 90 pos: 1 parser failed somehow near here: (EOL)
Sep 17 21:53:13 alarmpi lighttpd-angel[270]: lighttpd-angel.c.148: child (pid=273) exited normally with exitcode: 255

Works after I comment out the last line on the config file (include_shell "cat external.conf 2>/dev/null").
Not sure if it's a general issue, or just my system. (don't have another box to re-test)

max.bra commented on 2018-09-17 13:22 (UTC)

if someone is not happy with this page support there is another solution:

wipe hdd/sd/ssd/whatever, install raspbian/debian/ubuntu/fedora/centos and install official Pi-hole by

curl -sSL <> | bash

max.bra commented on 2018-09-07 10:09 (UTC)

api.php want to exec:

exec('sudo pihole disable');
exec("sudo pihole disable ".$disable."s");

exec('sudo pihole enable');

footer.js performs accordingly and respectively:



everything is terminal executed correctly.

there is no reason why something should not work. ideas?

max.bra commented on 2018-09-06 10:08 (UTC)

@graysky mmm, true. for me disable option is working but re-enable not, until FTL manual restart.

graysky commented on 2018-09-05 20:35 (UTC)

@max.bra - Does the "disable for 5 min" option work for you? It does not for me.

graysky commented on 2018-09-02 13:17 (UTC)

@max.bra - My bad... just took a long time to finally work. Great job with the 4.0 update btw!

max.bra commented on 2018-09-01 21:25 (UTC)

hi graysky, my "top clients (blocked only)" is a table!!! and there are reconds in it.

graysky commented on 2018-09-01 21:20 (UTC)

Is there a settings that would control the "top clients (blocked only)" section? After updating, the circular animated arrows just spin with no data.

max.bra commented on 2018-08-31 10:20 (UTC)

hi C0rn3j, update pi-hole-ftl first. obviously take a look at the wiki.

C0rn3j commented on 2018-08-31 10:02 (UTC) (edited on 2018-08-31 10:03 (UTC) by C0rn3j)

Can't update because of a dnsmasq conflict, what am I doing wrong?

pi-hole-server 3.3.1-1 > 4.0-8 ...

error: failed to commit transaction (conflicting files)

pi-hole-server: /etc/dnsmasq.conf exists in filesystem (owned by dnsmasq)

Errors occurred, no packages were upgraded.

max.bra commented on 2018-08-28 19:52 (UTC)

taking a note: "remember to update checksums after pkg change". it's on my monitor, i think i will never forget again...

uselessfire commented on 2018-08-28 18:59 (UTC)

Please update checksums in PKGBUILD after fix

max.bra commented on 2018-08-28 15:58 (UTC)

da... taken from upstream...

corrected, thanks for reporting.

ChuckHL commented on 2018-08-28 14:49 (UTC)

Need to fix your pihole-FTL.conf file comments. Regarding the valid values for BLOCKINGMODE you added the following in the comments:



Unfortunately that guide also needs to be fixed as IP-AAAA-NODATA is not a valid value (if you use it, it does nothing). The valid value is IP-NODATA-AAAA



max.bra commented on 2018-08-28 07:56 (UTC) (edited on 2018-08-28 08:24 (UTC) by max.bra)

@ChuckHL pihole error: sorry, wrong pihole error. yours is still present.

edit: no more and formatting

max.bra commented on 2018-08-28 07:46 (UTC)

yes it is!! thanks for digging it. updating wiki and basic ftl config file.

pihole user error: should be solved some pkg version ago.

ChuckHL commented on 2018-08-28 02:44 (UTC) (edited on 2018-08-28 03:08 (UTC) by ChuckHL)

Found the answer on this link.

and this:

The new blocking default method is redirecting everything blocked to

If you want the old behavior back, you need to add BLOCKINGMODE=IP OR BLOCKINGMODE=IP-NODATA-AAAA to /etc/pihole/pihole-FTL.conf

(Still not sure about the error I mentioned on my previous post about user 'pihole' not existing, and on the settings page under FTL Information, the User / Group shows as nobody / nobody)

ChuckHL commented on 2018-08-28 01:26 (UTC) (edited on 2018-08-28 02:23 (UTC) by ChuckHL)

I have uninstalled pi-hole-ftl and pi-hole-server and reinstalled but it did not help. One thing I notice thought when uninstalling is this error:

userdel: user 'pihole' does not exist error: command failed to execute correctly

When reinstalling the packages, I checked against /etc/passwd and no user pihole has been created.

P.S.: I am using pi-hole-server-4.0-8 and pi-hole-ftl-4.0-6 P.S.S: Another thing to mention is that my pi-hole box is not my DHCP server.

max.bra commented on 2018-08-27 18:45 (UTC)

Hi Taijian, no no everything is fine. ;-)

Taijian commented on 2018-08-27 18:40 (UTC)

One more thing (possibly?): I noticed that post_install we do:

chown -R http:http /srv/http/pihole

However, this is not done post_upgrade. Maybe it should? (no expert, just taking stabs in the twilight...)

max.bra commented on 2018-08-27 17:33 (UTC) (edited on 2018-08-27 17:34 (UTC) by max.bra)

how do i check? ps aux | grep dns

you should have stopped dnsmasq service before update and restart ftl after. as usual on any service replace update procedure. now it's late and is better and easier to reboot.

on my update pi-hole did not delete any of my 14 dns forward server... doh...

edit: formatting

max.bra commented on 2018-08-27 15:57 (UTC)

stupid question: have you definitely killed old dnsmasq and are you using FTLDNS only?

ChuckHL commented on 2018-08-27 15:48 (UTC) (edited on 2018-08-27 15:54 (UTC) by ChuckHL)

If I leave gravity.list in the new format (per se) without the host ip in front of each domain and I try to go to those websites, the browser does not know where to go and returns an error that could not resolve the name (ERR_NAME_NOT_RESOLVED). If I add the host in front of each line I now can see "website blocked" pi-hole page.

max.bra commented on 2018-08-27 15:44 (UTC)

it's intentional since dnsmasq integration:

only local.list is leaved in hosts format

if you browse, for example, you will receive tipical pihole blocking notification.

ChuckHL commented on 2018-08-27 15:37 (UTC) (edited on 2018-08-27 15:39 (UTC) by ChuckHL)

As a temp solution I am running "sed -i -e 's/^/ /' gravity.list" and "sed -i -e 's/^/ /' black.list" after pihole -g to fix the issue for now.

(of course is the ip of my pihole (for other people reading this) and needs to point to your own pihole)

max.bra commented on 2018-08-27 15:24 (UTC)

yes variable names are the same as before...

still looking...

ChuckHL commented on 2018-08-27 15:18 (UTC)

Checked SetupVars.conf and its the same as it was when I was running pihole v3. Not sure if IPV6 needs to be setup (I didnt have it on v3 either).


max.bra commented on 2018-08-27 15:15 (UTC)


same format for me. blocked traffic is so redirected to localhost. i don't know yet where the problem is. guessing it is on our SetupVars.conf...

looking into it...

ChuckHL commented on 2018-08-27 15:08 (UTC) (edited on 2018-08-27 15:09 (UTC) by ChuckHL)

Need help. Just updated from version 3 to 4 and now blocked sites are no longer redirecting to pihole server.

I noticed that my gravity.list (and also black.list) on version 3 showed results like this:

where as on version 4 is only like this:

I assume that is where the problem now starts since it no longer knows to redirect to pi-hole server's IP.

max.bra commented on 2018-08-27 08:33 (UTC)

hi jonandermb, "zero domain on blocklist" should be a problem about /etc/dnsmasq.conf: follow the arch pihole wiki.

DNS resolution not available should be this?

sounds strange you have this only on after update...

jonandermb commented on 2018-08-27 03:07 (UTC)

Just updated from 3.x to 4.0 and this is what I get if I want to update the lists: "[✗] DNS resolution is currently unavailable". I also have the "zero domains on blocklist" issue, although i see there's certainly blocking on my network.

phixion commented on 2018-08-26 16:11 (UTC) (edited on 2018-08-26 16:21 (UTC) by phixion)

I ran into another issue, it seems in the default config pihole is unable to write the sqlite databse file which means on every restart of the pi or the service your statistics are being reset. It's no showstopper but not nice either.

Dirty way to fix it is to change the permissions of the file/directory so the process can write to it.

I guess the elegant way is to have a dedicated serviceaccount for this and the right permissions.

edit: this probably belongs more to pi-hole-ftl than this package

max.bra commented on 2018-08-24 21:43 (UTC)

phixion commented on 2018-08-24 21:01 (UTC)

@domochevski that did the trick

domochevski commented on 2018-08-24 20:03 (UTC) (edited on 2018-08-24 20:04 (UTC) by domochevski)

I also had the "0 sites blocked" problem. To solve it: edit/write the /etc/dnsmasq.conf file and make sure you have "conf-dir=/etc/dnsmasq.d/,*.conf" inside it. Then restart the pi-hole-ftl service. Things should work now.

phixion commented on 2018-08-24 19:55 (UTC) (edited on 2018-08-24 19:56 (UTC) by phixion)

ran aurman -Rcs pi-hole-server pi-hole-ftl lighttpd

removed /etc/pihole|dnsmasq|lighttpd

systemctl stop|disable dnsmasq|pi-hole-gravity|pi-hole-ftl

rebooted and reinstalled with aurman -Syu pi-hole-server 4.0-2

still 0 blocking

max.bra commented on 2018-08-24 19:03 (UTC)

there are services to be stopped (dnsmasq) before uninstallation and services to be restarted (pi-hole-ftl). better and easier is to reboot.

Guernica commented on 2018-08-24 18:48 (UTC)

I also show "Domains on Blocklist" = 0, but /etc/pihole/gravity.list is not empty.

phixion commented on 2018-08-24 17:27 (UTC) (edited on 2018-08-24 17:34 (UTC) by phixion)


and click restart dnsmasq on webinterface is givin the same error

starting the service via systemd (lowercase 'ftl') works fine but its showing 0 addresses blocked everywhere for me.

johnnybash commented on 2018-08-24 17:24 (UTC)

got his problem. when changing DNS Servers for example on the webinterface and saving, the interface tries to restart the service. restart fails cause the service "pihole-FTL.service" doesn't exist

max.bra commented on 2018-08-24 17:09 (UTC)

@phixion hi, doesn't start what and from where? a little bit of debug precision can be of help...

johnnybash commented on 2018-08-24 16:44 (UTC)

works for me, but its pi-hole-ftl.service over here

phixion commented on 2018-08-24 16:41 (UTC) (edited on 2018-08-24 16:42 (UTC) by phixion)

Thanks for the update

doesn't start for me, though. looking into it now

[✗] Failed to restart pihole-FTL.service: Unit pihole-FTL.service not found.

SilverMight commented on 2018-08-21 04:06 (UTC)

don't mean to come off as impatient but any chance for 4.0 soon?

phixion commented on 2018-08-16 23:51 (UTC)

I see how it is! Enjoy your vacation.

max.bra commented on 2018-08-16 13:58 (UTC)

I was hoping to work on the packages from the beach but the WiFi is out of order... :-D

phixion commented on 2018-08-16 06:29 (UTC)

any chance to see this pushed to 4.0 anytime soon?

josealberto4444 commented on 2018-08-07 09:43 (UTC) (edited on 2018-08-07 10:50 (UTC) by josealberto4444)

Hi, tsago,

No, I get same error, it doesn't matter what I ask to dig:

$ dig +short @
dig: parse of /etc/resolv.conf failed
$ dig +short www.lkjasdfbo.buvoqwuyd @300.300.330.3331
dig: parse of /etc/resolv.conf failed

The thing is that it can actually resolve names, as I can ping anywhere:

$ ping
PING ( 56(84) bytes of data.
64 bytes from ( icmp_seq=1 ...

I'll ask in generic forums, thanks.

tsago commented on 2018-08-06 15:14 (UTC)

Hi, josealberto4444,

This looks more like a generic DNS issue, not really pi-hole related. Generic forums might be a better place for this.

But anyway, can you resolve something when you ask dnsmasq? E.g.: dig +short @ (or @

josealberto4444 commented on 2018-08-06 14:55 (UTC) (edited on 2018-08-06 15:00 (UTC) by josealberto4444)

Hi, I have a problem with resolving DNS requests from the machine that is actually running pi-hole as a server. I get this when trying to update lists:

$ pihole -g
[✗] DNS resolution is currently unavailable

Also, I think this is related, when using dig to find my external IP address:

$ dig +short        
dig: parse of /etc/resolv.conf failed

I looked for a solution in the pi-hole forums, but the most related thing was this one:
which links to these:
but all that was solved in v3.3 in theory.

Just in case, my /etc/resolv.conf:

$ cat /etc/resolv.conf
# This file is managed by man:systemd-resolved(8). Do not edit.
# This is a dynamic resolv.conf file for connecting local clients directly to
# all known uplink DNS servers. This file lists all configured search domains.
# Third party programs must not access this file directly, but only through the
# symlink at /etc/resolv.conf. To manage man:resolv.conf(5) in a different way,
# replace this symlink by a static file or a different symlink.
# See man:systemd-resolved.service(8) for details about the supported modes of
# operation for /etc/resolv.conf.

nameserver fe80::1%2

IP is the machine running pi-hole (Raspberry Pi 3B+).


max.bra commented on 2018-07-16 10:01 (UTC)


Exactly: this is not a normal pihole installation.

ASAP I will sed out that line. Thanks for reporting.

uint2048_t commented on 2018-07-16 04:21 (UTC) (edited on 2018-07-16 04:23 (UTC) by uint2048_t)

I've been noticing this line spamming the syslog every so often:

 Jul 15 18:29:27 zserve lighttpd-angel[9694]: sh: line 0: cd: /etc/.pihole/: No such file or directory

It looks like it's from line 204 in /srv/http/pihole/pihole/index.php

 $phVersion = exec("cd /etc/.pihole/ && git describe --long --tags");

On a 'normal' pihole install that directory would exist. Is it worth it to populate that variable with the pihole git version?

Taijian commented on 2018-07-12 11:25 (UTC)

OK, I was having a really weird issue with pi-hole and the fix I discovered was even more strange, but maybe you can help me figure out why this works this way?

So, my Odroid HC-1, running pi-hole and a Nextcloud stack would always fail to come back properly after a reboot because /etc/tmpfiles.d/pi-hole.conf would somehow not get executed and so there would be no dns and that would screw everything up. Now I experimented with various things and finally found that if I removed all the trailing spaces and '-' in that file, my problems would go away. So I guess I'm posting this both for awareness and because maybe someone could help me figure out why these were problematic?

graysky commented on 2018-06-21 18:49 (UTC)

@tholinka - It's probably an incorrectly configured web server. Recommend you uninstall that piece, reinstall it, and review the wiki for setup.

tholinka commented on 2018-06-21 08:06 (UTC) (edited on 2018-06-23 15:07 (UTC) by tholinka)

Hello, sorry if this has been posted before, but my google-fu is failing me.

I've been using this (and -ftl) fine for several months now, but then a couple weeks ago the graphs on the admin page quit loading. They sit spinning forever now (

Any advice on how to fix this?

I tried removing / reinstalling the packages to no avail.

max.bra commented on 2018-05-25 08:05 (UTC)

@gecko - we need log of FTL systemd service status when fail and errors about installing packages. Is it a file conflict before installing it or is an error post install running *.install script?

for me, the only pihole directory in /var is: /var/cache/lighttpd/compress/pihole and it's about admin gui web server.

gecko commented on 2018-05-25 07:56 (UTC) (edited on 2018-05-25 08:04 (UTC) by gecko)

@ max.bra Hi Max, thanks for your reply. What information can I give that would be more useful? I’d be happy to run some tests but I’d have to reinstall the package as I removed it. But I have never touched the curl command from the pi-hole site. This is a brand new installation of Arch Linux Arm on an R-Pi 3B+, using the aarch64 architecture, that I only installed last weekend. It was the first time I’d installed pi-hole on this machine (guess that also answers your question @graysky).

It’s interesting though, becuase there were certainly entries in my /var directory that are to do with pihole. I have already removed the package now and the directories have gone with it. But there was a pihole directory in /var/log that was owned by http, not the pihole user. So I thought that that might be part of the problem. I tried changing it but it made no difference.

--- edit

You know, I think I’ve gotten confused between pi-hole and pi-hole-ftl. I believe the problems I have been expereincing have been with FTL, not with pi-hole server itself. As, although the systemd pi-hole-ftl service was failing to start, I was able to resolve DNS using the server still. I think I got too caught up in the web interface not working (it was stuck on continuously loading) and thought that the whole thing was broken. I may reinstall the package and just use it from the CL.

--- edit

Also, I didn’t debug by using the pihole debug command, but by installing and using gdb. I was adapting the debian based instructions listed here:

graysky commented on 2018-05-24 22:05 (UTC)

@gecko - I have an ODROID-C2 running Arch ARM (aarch64) that runs pihole for me using these packages. Does your setup have something special you configured? These work out-of-the-box for me (aarch64 lxc).

max.bra commented on 2018-05-24 21:42 (UTC) (edited on 2018-05-24 21:49 (UTC) by max.bra)

@gecko hi gecko, unfortunately with the informations available I can not help. this package do not use /var directory at all for nothing. it is highly probable that you have also performed an official installation with curl from the pi-hole site. clean EVERYTHING thoroughly, install only the packages for arch and we'll talk about it again.

--- edit

and more: this package has debug run disabled/not available/deleted

--- edit 2

port 4711 has nothing to do with us

gecko commented on 2018-05-24 20:54 (UTC)

I can't run pi-hole-server from this package on my aarch64 installation. The systemd pi-hole-ftl service fails to start. Running pi-hole in debug mode, and I think it can't bind to the 4711 port. But more than that, I believe this package is not doing the correct permissions on /var/run/pihole. Either way, I can't run a new installation of pihole at all with this package :(

graysky commented on 2018-05-12 13:43 (UTC)

@tsago - All members of base-devel are implied deps for all AUR packages[1].


tsago commented on 2018-05-12 13:42 (UTC)

Hi crobe2,

Already discussed on e.g. 2017-08-02 in here. Please spend a few moments to check older comments next time. =>

(and if the link is not obvious enough, note that sudo belongs to base-devel)

crobe2 commented on 2018-05-12 13:35 (UTC)

The package is missing a dependency for "sudo", without it, many things, like the blocking page will not work.

tsago commented on 2018-05-03 15:58 (UTC)

FYI, they are >planning< to have a fork of dnsmasq in the near future =>

Note, however, that this change is yet to appear in the master branch. Currently it's still in testing. (so should not be present in the aur package yet)

max.bra commented on 2018-05-03 05:18 (UTC) (edited on 2018-05-07 06:15 (UTC) by max.bra)

hi lilmike, no no no. dnsmasq is always required and it's in official distribution too. FTL is not a dns server (nor proxy, nor wrapper, nor everything) and i think it will never be.

I read that pihole-ftl had an integrated dns server as of 3.0.

do you mean that: Add full DNSSEC support #186?

that is only for resolution.

and more: Switch to dnsmasq "extra" logging style to ensure 100% accuracy for the log analysis #174

dnsmasq is absolutely required.

lilmike commented on 2018-05-02 23:36 (UTC)

Hey, I read that pihole-ftl had an integrated dns server as of 3.0. We're now on pihole 3.3, but I can't find a way to enable it. It also says that enabling ftl will disable dnsmasq, but this doesn't happen for me. Is arch's package different? or is there a way I can enable ftl to do dnsmasq's work? I can't seem to find info how to enable it anywhere except in the release 3.0 post, and that doesn't seem to apply to the arch package at all. -Michael.

desertskunk commented on 2018-03-27 23:09 (UTC)

Sorry for the delay. Yeah, I cleared everything out and reran it and it seemed to work. Weird. Thanks for the quick response though!

max.bra commented on 2018-03-24 09:46 (UTC)

hi desertskunk, I've just checked and md5sum of both tarballs are the same. a momentary github problem?

desertskunk commented on 2018-03-24 08:16 (UTC)

I think the md5sum for has changed. I was trying to update to the most recent version and it's failing when trying to validate the source files for pihole-3.3.tar.gz.

artafinde commented on 2018-03-17 07:44 (UTC)

I think I'm experiencing the same behaviour as BIGFAT. My setup is similar also. LEDE server who manages the DHCP with advanced DNS options 6, The pi-hole server the purpose of just DNS server running on an ODROID c2 (archlinux). Works fine just doesn't show them in UI.

max.bra commented on 2018-03-16 19:19 (UTC) (edited on 2018-03-16 19:19 (UTC) by max.bra)


i can't see any relevant and recent changes here

nor here

max.bra commented on 2018-03-16 19:14 (UTC) (edited on 2018-03-16 19:15 (UTC) by max.bra)

@BIGFAT all seems normal...

That chart is a pre-rendered one drawed on a canvas with pixel fixed dimension. What happen if you collapse the width of your browser to 250/300 pixels? Do you see more destinations?

edit: wording

BIGFAT commented on 2018-03-14 17:10 (UTC)

Hi @max.bra, sorry i was busy the last few days. Here is a ~10min log in a pastebin: Im using LEDE with disabled rebind protection. The clients get the pihole IP addresses for ipv4 and v6 directly trough the router interface settings (for example: 6,192.168.x.x). Dnsmasq on LEDE resolve only for the device itself. Like i said everything run fine, beside the webpage not showing all dns resolvers under "Forward Destinations". It did the release before. It must have something to do with recent changes.

max.bra commented on 2018-03-01 06:59 (UTC) (edited on 2018-03-01 07:00 (UTC) by max.bra)

Hi Flat, gravity_CheckDNSResolutionAvailable() look for pi.hole as a lookup domain. dnsmasq (via addn-hosts=/etc/pihole/local.list) is awared of that hostname. if you don't use localhost as resolver you need to update your /etc/hosts with pi.hole

Flat commented on 2018-03-01 01:43 (UTC)

This bug doesn't appear to be fixed. I've setup pihole according to the wiki and the blocklists would not update until I set /etc/resolv.conf to, can anyone else confirm this bug is still an issue for at least arch?

max.bra commented on 2018-02-27 07:52 (UTC)

@BIGFAT Hi Bigfat, sorry for late replay. Not using that option, can you please pastebin an excerpt of your /run/log/pihole/pihole.log, especially theforward lines? And what is your router?

BIGFAT commented on 2018-02-26 19:07 (UTC)

Hi, i have a small problem: Im using several dns server with the "all-servers" option to augment the performance. But the gui doesnt show any of them beside local under "Forward Destinations". Pihole logs show me that all dns get queries and beside that, pihole is working without any issue. Any Idea how i can fix this cosmetic "bug"? Greeting and thank you for the good work.

ChuckHL commented on 2018-02-24 05:19 (UTC) (edited on 2018-02-24 05:29 (UTC) by ChuckHL)

First of all, thanks for your updates and keeping this app updated (i really love pi-hole).

Edit: I asked about two errors that you had already addressed in the installation instructions.

Thanks for everything. I love this app. Thumbs up.

max.bra commented on 2018-02-16 15:37 (UTC)

@ryan updated! thanks for reporting.

ryanpetrello commented on 2018-02-16 14:50 (UTC) (edited on 2018-02-16 14:50 (UTC) by ryanpetrello)


Looks like a change in dnsmasq log format requirements:

Configuring dnsmasq with log-queries=extra 100% solved the problem for me. Maybe we could update the wiki?

max.bra commented on 2018-02-16 14:33 (UTC)

@ryanpetrello my dnsmasq log format is standard and the id is parsed correctly

are you on arm or x86_64? is your archlinux official? what's in your locale.conf? 'locale' output?

ryanpetrello commented on 2018-02-16 14:05 (UTC) (edited on 2018-02-16 14:13 (UTC) by ryanpetrello)


This indeed looks like an FTL bug. I tried building my own with this patched version of parser.c:

After making this change, it works - I can see queries in the SQLite database and in the web UI.

Maybe the format of my dnsmasq changed somehow?

ryanpetrello commented on 2018-02-16 13:09 (UTC) (edited on 2018-02-16 13:14 (UTC) by ryanpetrello)

@max.bra very recent arch install (last few days), and I installed from AUR. Here's a snippet of pihole-FTL.log immediately after a systemctl restart:

It's definitely answering DNS, it's just that the database file is never written to (and so the admin console shows nothing).

Perhaps those Error getting ID for query are related?

max.bra commented on 2018-02-16 12:53 (UTC)

hey ryanpetrello, a little bit more about your system and install method: Is your arch clean and did you install from AUR packages only? Can you pastebin your FTL log? (/run/log/pihole-ftl/pihole-FTL.log)

ryanpetrello commented on 2018-02-16 12:40 (UTC) (edited on 2018-02-16 12:41 (UTC) by ryanpetrello)

@max.bra @graysky I've got pi-hole-ftl installed and it's properly blocking ads, but the statistics database never gets written to, so I can't see statistics:

~ dig | grep amazon ; <<>> DiG 9.8.3-P1 <<>> ; IN A 300 IN A (ip-of-pi.hole)

~ sudo grep DBFILE /etc/pihole/pihole-FTL.conf


~ ls -la /etc/pihole/pihole-FTL.db

-rw-r--r-- 1 pihole pihole 20480 Feb 16 07:21 /etc/pihole/pihole-FTL.db

~ sudo sqlite3 /etc/pihole/pihole-FTL.db

SQLite version 3.22.0 2018-01-22 18:45:57

Enter ".help" for usage hints.

sqlite> SELECT COUNT(*) FROM queries;


Any ideas or leads?

max.bra commented on 2018-02-09 16:45 (UTC) (edited on 2019-10-18 23:14 (UTC) by max.bra)

ArchLinux Pi-hole is not officially supported by Pi-hole project. In case of bugs and malfunctions please DO NOT file a report upstream.

First of all check if the wiki ( can help then ask here for assistance and tips.
When it will be excluded that the problem does not depend on ArchLinux we will file a bug upstream.

max.bra commented on 2018-01-24 09:45 (UTC)

Hi Taijian and thanks for reporting. You are perfectly right! synced with upstream.

Taijian commented on 2018-01-23 22:42 (UTC)

I have the problem that my pi-hole resets the FTL output at midnight of every day, even though I have the config option 'TIMEFRAME=rolling24h' set in my pihole-FTL.conf. Investigating, I found that upstream calls pihole flush with the parameter 'once' when doing the nightly flushing, specifically so that the data of the previous day is still visible to FTL. Leaving out this parameter causes logrotate to be called twice. See here:

I would therefore propose to change pi-hole-logtruncate.service as follows:

ExecStart=/usr/bin/pihole flush ONCE

max.bra commented on 2017-12-31 18:07 (UTC)

hi MattScherbatsky, thanks for reporting and yes, v. 3.3 it's now a release candidate. maybe an upstream release error. package is now on 3.2.1, force update if you installed v. 3.3

MattScherbatsky commented on 2017-12-31 16:04 (UTC) (edited on 2017-12-31 16:12 (UTC) by MattScherbatsky) 404: Not found.

Is there an available mirror?

EDIT: it seems that the latest version is 3.2.1. Changed it, use this md5sum: a0ff29ba87d22b695baa48194b1ebfd2 and it works

max.bra commented on 2017-12-11 22:13 (UTC)

done, thank you!

graysky commented on 2017-12-11 22:11 (UTC)

@max.bra - You introduced lighttpd.conf with the 3.2-2 commit but you are installing lighttpd.pihole.conf in the PKGBUILD. That first one seems to be unused. They are different from each other.

max.bra commented on 2017-12-11 21:52 (UTC)

@graysky removed -R, my fault...

max.bra commented on 2017-12-11 21:49 (UTC) (edited on 2017-12-11 21:49 (UTC) by max.bra)

@graysky there is no need to recourse all, all other files are sudoed accessed. our /etc/pihole is 755, i know. but my /etc/pihole is always 777 and i don't know why. i have never found a chmod 777 in the project... bah...

graysky commented on 2017-12-11 21:45 (UTC)

@max.bra - I haven't updated to 3.2-2 yet but simply chown pihole:pihole /etc/pihole was all that's needed... the -R switch will make all files under there pihole:pihole when it seems that they should be root:root. Also, my /etc/pihole is 755 not 777.

max.bra commented on 2017-12-11 21:41 (UTC)

@MartB thanks for reporting, patch applied. Changed /etc/pihole ownership too but, as for now, my /etc/pihole directory is always chmodded to 777 by the running pihole...

MartB commented on 2017-12-11 20:18 (UTC) (edited on 2017-12-11 20:34 (UTC) by MartB)

@schmodd apply this patch right after the version.patch file in the pkgbuild.

This will be in the next version for sure and it fixes your issue if you install bind-tools

schmodd commented on 2017-12-11 17:32 (UTC) (edited on 2017-12-11 18:47 (UTC) by schmodd)

Hi, it seems that Pihole 3.2 is working fine so far with one exception: When I try to update my lists (via webinterface - tools) I get:

[✗] DNS resolution is currently unavailable

[✗] DNS resolution is not available

Ok, they got it on the buglist on github already, nothing to do with the package here i guess. SRY.

artafinde commented on 2017-12-09 10:31 (UTC)

@max.bra ping if you need help testing or packaging.

max.bra commented on 2017-12-09 09:50 (UTC)

Hi batal, 3.2 is a big release. I think that the package should be ready for beginning of he week.

batal commented on 2017-12-09 08:35 (UTC)

Hello! I don't want to press or anything, just am excited: Do you have a schedule for version 3.2 yet? Thanks for maintaining the package. I couldn't use Arch if you would not :-)

max.bra commented on 2017-12-03 20:14 (UTC)

hi felis, thanks for reporting.

commented on 2017-12-03 17:38 (UTC)

Please add iproute2 to the dependencies since the initial block list generation depends on it.

steinpfeffer commented on 2017-11-25 15:27 (UTC)

@max.bra: I do not want to make such a fundamental change to the running system. Thus I gave up - unfortunately.

max.bra commented on 2017-11-25 13:01 (UTC)

@steinpfeffer have you been able to install Pi-hole?

max.bra commented on 2017-11-22 20:47 (UTC)

mmm, you should make it with a mount point

steinpfeffer commented on 2017-11-22 19:12 (UTC)

Hi Max, not if I made it a symlink because the data resides on another hard drive ...

max.bra commented on 2017-11-22 19:00 (UTC)

Hi steinpfeffer, /srv is not a symlink and is provided by filesystem package...

steinpfeffer commented on 2017-11-22 18:40 (UTC) (edited on 2017-11-22 18:40 (UTC) by steinpfeffer)

I have problems installing the package. It says there are conflicting files because /srv already exists. /srv is a symlink. Any ideas?

artafinde commented on 2017-11-13 23:35 (UTC)

@max.bra works now - thanks.

max.bra commented on 2017-11-13 09:03 (UTC)

@artafinde new revision: can you please update and retry with: # /opt/pihole/ $ gravity -g $ awk -F " " '{print $1}' /etc/pihole/gravity.list | sort | uniq

artafinde commented on 2017-11-12 10:18 (UTC)

@max.bra Thanks but that didn't work either $ cat /etc/pihole/setupVars.conf PIHOLE_INTERFACE=eth0 IPV4_ADDRESS= IPV6_ADDRESS=

max.bra commented on 2017-11-12 10:02 (UTC) (edited on 2017-11-12 10:04 (UTC) by max.bra)

@artafinde what an idiot, my fault, little forget. your IPv6 addresses now seems to be good but Pi-hole still doesn't know. # /opt/pihole/ $ gravity -g $ awk -F " " '{print $1}' /etc/pihole/gravity.list | sort | uniq now you should see both addresses type and inside /etc/pihole/setupVars.conf, IPV6_ADDRESS= entry should be valued. mmm, wiki stuff

artafinde commented on 2017-11-12 09:45 (UTC) (edited on 2017-11-12 09:58 (UTC) by artafinde)

@max.bra after `pihole -g` I checked the list and it has only IPv4 $ awk -F " " '{print $1}' /etc/pihole/gravity.list | sort | uniq The admin interface is showing empty:

max.bra commented on 2017-11-11 15:56 (UTC)

@artafinde > pi-hole can't detect it. why did you say that? please execute 'pihole -g' and see if inside /etc/pihole/gravity.list you can find both your IPv4 and IPv6 Pi-hole box addresses.

artafinde commented on 2017-11-11 15:28 (UTC)

@max.bra hmm I didn't know about the Link local address thanks for that. I've reconfigured the DHCP to provide IPv6 (I think) but still pi-hole can't detect it. $ ip -6 a show eth0 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qlen 1000 inet6 fded:50d6:9e76::ce6/128 scope global noprefixroute valid_lft forever preferred_lft forever inet6 fded:50d6:9e76:0:21e:6ff:fe33:e942/64 scope global mngtmpaddr noprefixroute valid_lft forever preferred_lft forever inet6 fe80::21e:6ff:fe33:e942/64 scope link valid_lft forever preferred_lft forever It's not a biggie just trying to understand it.

max.bra commented on 2017-11-11 14:54 (UTC)

Hi artafinde, yes, Pi-hole is IPv6 capable on an full IPv6 lan configuration. Your IP is a link local type one ( and IPv6 capabilities are limited to internet DNS resolution.

artafinde commented on 2017-11-11 11:25 (UTC)

Is this IPv6 capable? My admin page has no IPv6 detected even there is an IPv6 assigned on the pi-hole host. $ ip -6 a list eth0 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qlen 1000 inet6 fe80::21e:6ff:fe33:e942/64 scope link valid_lft forever preferred_lft forever

max.bra commented on 2017-11-02 11:25 (UTC)

v. 3.1.4-8 - for those using nginx it would be wise to update its config

max.bra commented on 2017-10-13 11:24 (UTC) (edited on 2017-10-13 11:24 (UTC) by max.bra)

v. 3.1.4-6 - nginx config changed edit: typo

max.bra commented on 2017-10-09 18:33 (UTC)

Hey BIGFAT! well, you can still use the old config file, no problem at all. or, you can follow the always excellent arch wiki:

BIGFAT commented on 2017-10-09 18:29 (UTC)

Hallo max.bra. Since im a beginner, it would be nice if you can tell me how. I dont know where to put exactly the list. Thank you.

max.bra commented on 2017-10-09 15:55 (UTC)

since it is almost a puddle, i've removed open_basedir directive from lighttpd config example. pihole needed dirs list is now on the wiki for a rapid personal integration. cheers.

max.bra commented on 2017-10-05 20:14 (UTC) (edited on 2017-10-05 20:15 (UTC) by max.bra)

as usual, open_basedir directory missing. done and thanks for reporting. be careful, update your lighttpd.conf

tsago commented on 2017-10-05 19:38 (UTC) (edited on 2017-10-05 20:03 (UTC) by tsago)

I can reproduce BIGFAT's issue on a >fresh< Arch VM, running latest pi-hole-server 3.1.4-3. Using lighttpd for pihole web-interface. Followed setup guide from arch wiki. GUI -> Settings -> Teleporter -> click on Export Same error as BIGFAT posted. Error comes from script /srv/http/pihole/admin/scripts/pi-hole/php/teleporter.phpm line 170. ( ) Sounds like the "http" service user cannot create the temporary zip file? I know nothing about PHP, so I'm hitting a wall here... edit: works when I change the folder from /tmp to /srv/http/pihole (line 164 of teleporter.php), so I'm pretty sure it's write permissions. Broken setup on my side? Arch-specific pi-hole issue? General pi-hole issue that should be reported upstream? (I guess they'd notice already if this was issue from upstream)

graysky commented on 2017-10-05 18:40 (UTC)

No issue using nginx here. You haven't provided any level of detail for someone to help you. Also, the AUR isn't the appropriate place for this. Recommend you open a thread on the bbs and provide some details beyond "I can't use it."

BIGFAT commented on 2017-10-05 16:05 (UTC) (edited on 2017-10-06 17:01 (UTC) by BIGFAT)

Hi, i have a small problem. I cant use the pi-Hole teleporter: cannot open/create Error: PHP user: http Thank you for the good work. PS: Sorry for not giving details, but yes it was on lighttpd.

max.bra commented on 2017-09-10 19:44 (UTC)

hi tdlk, especially with the latest package modifications after the arrival of graysky (now the package is freer and less automated, leaving the user many more decisions and configurations) it is not mandatory to use the conf recommended for lighttpd, you can use your config whenever you want. Inserting open_basedir (in the original project there is no open_basedir directive) wants to give you a little more security handled directly by the maintainers. If one has different needs of course, he must take different paths.

tdkl commented on 2017-09-10 17:46 (UTC) (edited on 2017-09-10 18:09 (UTC) by tdkl)

Can anyone explain me a good reason why open_basedir is defined in lighttpd.conf? Because while it might be fine if you only run pi-hole, it's a PITA if you want to run something else, for example Mediawiki on /mediawiki. Leaving the open_basedir out of lighttpd.conf and entering it in php.ini works for both. This is the main problem I started to have with pi-hole a while ago - it's assuming it's the only player on ones system and not being a good citizen. It'll probably replace systemd at some point.

max.bra commented on 2017-08-02 13:21 (UTC)

hi ahrs, no should not. $ pacman -Qs sudo local/sudo 1.8.20.p2-1 (base-devel)

ahrs commented on 2017-08-02 05:31 (UTC) (edited on 2017-08-02 05:34 (UTC) by ahrs)

sudo should be added as a dependency since it's used to display the Status in the web interface (

graysky commented on 2017-07-22 15:18 (UTC)

I haven't used it as a DHCP before. You might wanna wait for max to comment. If not, report upstream. Do know that the Arch package does modified their files as you see in the PKGBUILD so it's possibly something we're doing. The best test is for you to spin up an ubuntu lxc and try it via their script. If it repeats, they're to blame :)

tsago commented on 2017-07-22 14:59 (UTC)

@graysky: Indeed. Pi-hole is the DHCP server. Config almost default (clean reinstall recently), only added DHCP range, local domain name, and manually imported static leases. (Note: tested with empty/no static leases file too) No improvement after restarting ftl or dnsmasq. The IP of the static lease is out of the DHCP range, though I don't think that should matter.

graysky commented on 2017-07-22 14:45 (UTC)

@tsago - That is using pi-hole as a DHCP server I assume?

tsago commented on 2017-07-22 14:31 (UTC)

Before I start blaming upstream, anybody observing this issue as well? Web-interface does not show any static leases. However, it successfully adds new static leases (I see new additions in /etc/dnsmasq.d/04-pihole-static-dhcp.conf). Several times even, if I retry the same static lease. Seems to be an issue purely with reading or displaying the static leases by the web interface.

graysky commented on 2017-07-01 19:41 (UTC)

Wild guess: does /usr/bin/makepkg honor aliases setup by users? I can't check the source now to see.

max.bra commented on 2017-07-01 15:08 (UTC)

any sed guru in the house? in ChuckHL configuration (and only for him, right now) in PKGBUILD prepare section, a sed seems to fail: # change log and bin location to logrotate sed -i "s|/var/log/pihole.log|/run/log/pihole/pihole.log|w $_ssc" "$srcdir"/$_pkgname-$pkgver/advanced/logrotate if [ -s $_ssc ] ; then rm $_ssc ; else echo " ==> Sed error: change log and bin location to logrotate 1" && return 1 ; fi sed -i "s|/var/log/pihole-FTL.log|/run/log/pihole-ftl/pihole-FTL.log|w $_ssc" "$srcdir"/$_pkgname-$pkgver/advanced/logrotate if [ -s $_ssc ] ; then rm $_ssc ; else echo " ==> Sed error: change log and bin location to logrotate 2" && return 1 ; fi first is successful, second fails. Moreover, if not able to substitute requested string second sed should fail and abort package creation...

max.bra commented on 2017-07-01 14:45 (UTC)

mmm, here it is let me try to understand why...

ChuckHL commented on 2017-07-01 14:26 (UTC) (edited on 2017-07-01 14:36 (UTC) by ChuckHL)

I completely reinstalled everything (piholeserver and ftl). I can see both directories: [root@alarm pihole]# ls -lah /run/log/ drwxr-xr-x 2 dnsmasq root 160 Jul 1 00:00 pihole drwxr-xr-x 2 pihole pihole 60 Jan 31 19:09 pihole-ftl [root@alarm pihole]# ls -lah /var/run/log/ drwxr-xr-x 2 dnsmasq root 160 Jul 1 00:00 pihole drwxr-xr-x 2 pihole pihole 60 Jan 31 19:09 pihole-ftl I still have the error. The interesting thing is that logrotate from the latest version on my devices seems different than yours. Here is the output of logrotate on the latest version on my machine: [root@alarm pihole]# cat logrotate /run/log/pihole/pihole.log { ... } /var/log/pihole-FTL.log { ...} Edit: I just modified logrotate I changed: /var/log/pihole-FTL.log { to this: /run/log/pihole-ftl/pihole-FTL.log { And finally, the problem was fixed.

max.bra commented on 2017-07-01 10:43 (UTC) (edited on 2017-07-01 10:43 (UTC) by max.bra)

this is current logrotate contents: $ cat /etc/pihole/logrotate /run/log/pihole/pihole.log { [...] } /run/log/pihole-ftl/pihole-FTL.log { [...] } and this is the situation since it was released FTL. Nothing point to /var, at least for now. have you ever executed official shell script ''?

ChuckHL commented on 2017-06-30 21:31 (UTC) (edited on 2017-06-30 21:38 (UTC) by ChuckHL)

Not sure if I typed it correctly or on what directory I have to be in when running it. It gave me no output. [root@alarm ~]# grep -i "var/log" /opt/pihole/* /usr/bin/pihole [root@alarm ~]# grep -i "/var/log" /opt/pihole/* /usr/bin/pihole [root@alarm ~]#

max.bra commented on 2017-06-30 21:01 (UTC)

this is no sense!! please can you output me this: grep -i "var/log" /opt/pihole/* /usr/bin/pihole

ChuckHL commented on 2017-06-30 20:53 (UTC) (edited on 2017-06-30 21:36 (UTC) by ChuckHL)

First of all, I had not updated pihole-ftl to the latest, so my bad on that. I was running pihole-ftl 2.9.1-1 rather than 2.9.4-3. Also, I just updated piholeserver from 3.1-7 to 3.1-8. I just ran pihole -l off and I have the same issue. [root@alarm pihole]# pihole -l off ::: Flushing /run/log/pihole/pihole.log ...error: stat of /var/log/pihole-FTL.log failed: No such file or directory error: stat of /var/log/pihole-FTL.log failed: No such file or directory ... done! ::: Logging has been disabled! [root@alarm pihole]# pihole -l on ::: Logging has been enabled! [root@alarm pihole]# [root@alarm pihole]# pihole -v Pi-hole version is 3.1 (Latest: v3.1) AdminLTE version is 3.1 (Latest: v3.1) FTL version is v2.9.4 (Latest: v2.9.4) [root@alarm ~]# pacman -Qs pi-hole local/pi-hole-ftl 2.9.4-3 The Pi-hole FTL engine local/pi-hole-server 3.1-8 The Pi-hole is an advertising-aware DNS/Web server. Arch adaptation for lan wide DNS server. [root@alarm pihole]# uname -a Linux alarm 4.11.7-1-ARCH #1 SMP Sat Jun 24 19:53:12 MDT 2017 aarch64 GNU/Linux

max.bra commented on 2017-06-30 07:23 (UTC) (edited on 2017-06-30 07:28 (UTC) by max.bra)

all normal here: $ pihole -l off ::: Flushing /run/log/pihole/pihole.log ...... done! ::: Logging has been disabled! $ pihole -l on ::: Logging has been enabled! $ pihole -v Pi-hole version is 3.1 (Latest: v3.1) AdminLTE version is 3.1 (Latest: v3.1) FTL version is v2.9.4 (Latest: v2.9.4) $ pacman -Qs pi-hole local/pi-hole-ftl 2.9.4-3 local/pi-hole-server 3.1-7

ChuckHL commented on 2017-06-29 23:40 (UTC)

Another possible bug: [root@alarm ~]# pihole -l off ::: Flushing /run/log/pihole/pihole.log ...error: stat of /var/log/pihole-FTL.log failed: No such file or directory error: stat of /var/log/pihole-FTL.log failed: No such file or directory ... done! ::: Logging has been disabled! If I recall correctly arent logs stored at /var/run/log/pihole rather than /var/log/pihole or /run/log/pihole?

max.bra commented on 2017-06-29 20:45 (UTC) (edited on 2017-06-29 21:04 (UTC) by max.bra)

oh lord!! those commands should not be there!! pihole update must never be executed, yes. debug is theoretically working but not supported yet. patching immediately... edit2: ok done. help output cleaned up. thanks for reporting!! edit: ah ok, it's only a matter of help output. all debug and update commands are sedded away.

ChuckHL commented on 2017-06-29 14:33 (UTC) (edited on 2017-06-29 14:37 (UTC) by ChuckHL)

Hi @max.bra First of all. Thank you for your time maintaining this pkg. It's one of my most used defences @ my home network a long with privoxy and a Asus router (using tomato system with iptables to force all devices to use pihole and privoxy). I have a question. [root@alarm ~]# pihole ... Debugging Options: -d, debug Start a debugging session Add '-a' to enable automated debugging ... Options: ... -up, updatePihole Update Pi-hole subsystems ... [root@alarm ~]# Are those options, debug and updatepihole supposed to work? I suppose update should not as it would point to the original pihole version which is not cooked for archlinux. But how do I use debug? Regards

max.bra commented on 2017-06-29 07:44 (UTC)

@ChuckHL yes! you are perfectly right. my fault.

ChuckHL commented on 2017-06-29 01:01 (UTC)

I found a possible bug on the latest version 3.1-5. It seems that any domain added to the blacklist is not being blacklisted and is ignored. The problem seems to be that "pihole -g" (or update from the website) no longer pulls the blacklist domains into the gravity.list but rather creates a black.list file which is not added to /etc/dnsmasq.d/01-pihole.conf file.

max.bra commented on 2017-06-28 21:08 (UTC) (edited on 2017-06-29 08:29 (UTC) by max.bra)

@ChuckHL that folder is part of pi-hole git repo since forever but nothing point to it in installation shell script: or is something atavic or something that will be implemented later on...

graysky commented on 2017-06-28 20:59 (UTC)

Does browsing to http://pi.hole/admin not work for others?

ChuckHL commented on 2017-06-28 04:56 (UTC)

I notice that other pihole installation (outside of ArchLinux) create a folder named /etc/.pihole that includes some directory named "block hulu ads" and has some extra files for blocking such ads. Will you include that in the following releases (or is it outside of the scope for you)?

max.bra commented on 2017-06-24 23:07 (UTC) (edited on 2017-06-24 23:44 (UTC) by max.bra)

hi bofslime, thanks for reporting. > sqlite3.c:84551:1: note: here > case OP_AggStep: { > ^~~~ this is pi-hole-FTL compilation: messages like those are compilation warnings and you can ignore them. your system become unresponsive because FTL compilation is rather heavy for an arm processor. it's absolutely normal especially for a monocore cpu. not so normal the error at the end. ChuckHL too was reporting similar compilation error. try to download ftl package snapshot and recompile it or try to crosscompile it ( with a more powerfull machine.

bofslime commented on 2017-06-24 22:41 (UTC)

Seeing frequent issues when updating pi-hole from the AUR. My system will often become unresponsive when in this state and either have to wait it out, or preemptively have a 2nd session open as it stops accepting new sessions when stalling on building. Often this includes negatively affecting DNS resolution. sqlite3.c:84548:15: warning: this statement may fall through [-Wimplicit-fallthrough=] pOp->opcode = OP_AggStep; ^ sqlite3.c:84551:1: note: here case OP_AggStep: { ^~~~ {standard input}: Assembler messages: {standard input}:147265: Warning: end of file not at end of a line; newline inserted {standard input}:147690: Error: ']' expected -- `ldr r3,[fp,#-1' gcc: internal compiler error: Killed (program cc1) Please submit a full bug report, with preprocessed source if appropriate. See <> for instructions. make: *** [Makefile:57: obj/sqlite3.o] Error 4 ==> ERROR: A failure occurred in build(). Aborting... The build failed.

max.bra commented on 2017-06-21 06:44 (UTC)

@graysky > Shall we create a wiki page for this package? Created a stub yesterday evening for server and standalone too > recommend you pass the version on to the source: yes nice shot!

graysky commented on 2017-06-20 22:26 (UTC)

@max.bra - recommend you pass the version on to the source: source=(pihole-$pkgver.tar.gz::$_pkgname/$_pkgname/archive/v$pkgver.tar.gz admin-$_wwwpkgver.tar.gz::$_pkgname/$_wwwpkgname/archive/v$_wwwpkgver.tar.gz

graysky commented on 2017-06-20 21:43 (UTC)

@max.bra - I have nginx-mainline working with this package. I will try to write up the steps. Shall we create a wiki page for this package?

graysky commented on 2017-06-20 19:26 (UTC)

@max.bra - I think the best method here is simply to put them in the optdepends array and to provide a template config for each. As well, you give some post-install instructions in /usr/share/doc/pihole/configuration which is the best place in my opinion for this. I recommend that you remove the systemctl calls from the .install files for this and the ftl package. Put the instructions in the configuration docs but never have pacman call systemctl unless data loss will result if you don't do it. I am pretty busy with life right now but I will do my best to help with the documentation and a nginx example config. I just can't devote lots of time to it in the short term.

moonman14 commented on 2017-06-20 19:09 (UTC)

@max.bra I had previously made the change but I had not restarted yet. After restarting my pi everything worked as expected. Thanks for your help.

max.bra commented on 2017-06-20 18:18 (UTC)

moonman14: have you followed the pinned comment of this page? *** i know, i know. it's time for a wiki... *** :facepalm

moonman14 commented on 2017-06-20 17:14 (UTC)

@max.bra No, still not working for me.

max.bra commented on 2017-06-20 16:33 (UTC) (edited on 2017-06-20 16:33 (UTC) by max.bra)

@graysky @moonman14 Have you identified the inconsistency with the instructions?

max.bra commented on 2017-06-20 16:28 (UTC)

@graysky > I also recommend creating a page on the Arch Wiki to help. yes, i know. this is only a matter of time i wrote lots of time that i'm available to integrate an nginx config. if someone that know nginx can provide me a template, it's not a problem for me to create an integration or a pi-hole-server-nginx too (not the best way, maybe a multi choice package install file is better)

graysky commented on 2017-06-20 14:44 (UTC)

@moonman14 - Same problems for me on odroid or x86_64

graysky commented on 2017-06-20 14:21 (UTC)

@max.bra - Not sure what the best way is for it, but certainly, it shouldn't require any specific http server as many will work. Perhaps provide a few config files and leave it to the user to select one. I also recommend creating a page on the Arch Wiki to help.

moonman14 commented on 2017-06-20 13:46 (UTC)

Hello, I am having an issue with the initial configuration. I followed your instructions and all of the required services are running. However the graphs on the web interface do not load and dns queries are not answered. This problem is occurring on my desktop as well as my Pi Zero so I assume that I am making a mistake with the configuration.

max.bra commented on 2017-06-20 12:21 (UTC)

How could it be done? Settings lighttpd optional? Does it seems a good idea?

graysky commented on 2017-06-20 11:54 (UTC)

Since alternative servers can be use (ie nginx) are you able to adjust the depends array to allow for this?

max.bra commented on 2017-05-14 13:57 (UTC)

@ChuckHL @Vrob sorry guys, i forgot about it. solved.

Vrob commented on 2017-05-14 12:26 (UTC)

Can confirm, I'm having the same logrotate issue.

ChuckHL commented on 2017-05-09 02:49 (UTC) (edited on 2017-05-09 03:33 (UTC) by ChuckHL)

I'm having a problem flushing the logs (located in /run/log/). When I use the command on terminal of "pihole -l off" it does not delete the log. And flushing the log from the web interface is not working either. Error on screen: ::: Flushing /run/log/pihole/pihole.log ...error: cannot stat /etc/pihole/logrotate: No such file or directory Had to reinstall the previous version of pihole to retrieve logrotate from it, and then reinstall pihole latest version.

max.bra commented on 2017-05-07 17:18 (UTC)

Pi-hole 3.0.1 and PHP admin 3.0.1a - it's needed to enable php sockets extension: # nano /etc/php/php.ini ; ===> - graphs malformed on google chrome

Ubermensch commented on 2017-05-06 20:34 (UTC)

The version 3.0.1 is out please update :)

max.bra commented on 2017-05-04 17:32 (UTC)

hi Splith, sudo is part of base-devel group that is a prerequisite for AUR building.

Splith commented on 2017-05-04 17:19 (UTC)

This package requires sudo but it isn't in the dependencies list, can you add it?

max.bra commented on 2017-04-10 16:20 (UTC) (edited on 2017-04-10 16:21 (UTC) by max.bra)

hi geekinthesticks, those commands are not supported in archlinux pihole distribution and are just removed by PKGBUILD prepare script. you should not even see them. --- Original pi-hole install by a script (on supported distro, not arch). We install via AUR package. Too different to support reconfiguring. you have to simply reinstall the package. ::: -r, reconfigure was an oversight now resolved in the latest revision (thanks for reporting) ---- debug is maybe feasible, one of these days.

geekinthesticks commented on 2017-04-10 15:58 (UTC) (edited on 2017-04-10 17:44 (UTC) by geekinthesticks)

Thanks. It's easy enough to edit the files manually in any case.

max.bra commented on 2017-02-04 11:03 (UTC) (edited on 2017-02-05 22:04 (UTC) by max.bra)

hi tsago, thanks for reporting. done. > Also, before I flushed, the graph in GUI was stuck on always showing the same day. if i have well understood the observation: the dataset used for drawing that canvas collects data from pihole.log only and should show the time lapse between the first log timestamp and the last one. there is no "today filter" in the code... edit: typo

tsago commented on 2017-02-04 09:35 (UTC)

Hi there! Looks like flushing with logrotate is broken. # pihole -f ::: Flushing /run/log/pihole/pihole.log ...error: cannot stat /etc/pihole/logrotate: No such file or directory ... done! Fixed by adding into /etc/pihole/logrotate with path modified to /run/log/pihole/pihole.log. Also, before I flushed, the graph in GUI was stuck on always showing the same day. Did I do something wrong? Or is it arch-pihole specific? Or is it a general issue with the last version?

max.bra commented on 2017-01-19 08:19 (UTC)

Hi Chilloutman, yes it is. Missed that. Thanks for reporting.

Chilloutman commented on 2017-01-19 07:14 (UTC)

During the installation "netstat" is used and on my system this caused some issues as it was not available. After installing "net-tools" from core, the errors were fixed. Is "net-tools" a missing dependency?

tdkl commented on 2017-01-14 21:42 (UTC)

@max.bra: no I use lighttpd @C0rn3j: if you mean the js console inside the browser, there were none. I've installed 2.9.5 back from AUR, works without issues. Too much hassle at constant changes for each update, so I'll keep it at this version.

C0rn3j commented on 2017-01-12 10:29 (UTC)

@tdkl Check the javascript console for errors

max.bra commented on 2017-01-12 07:31 (UTC)

It seemed like a good start :-P

C0rn3j commented on 2017-01-11 23:57 (UTC)

Taillog is fine for me over HTTPS, so it must be a different issue ^^

max.bra commented on 2017-01-11 23:37 (UTC)

hi tdkl, all functional here. if I remember correctly, are you using nginx? maybe via https:

tdkl commented on 2017-01-11 15:55 (UTC)

I've noticed that the query logs aren't displayed anymore. Am I only one with this issue ? I've reinstalled pihole freshly from AUR, but there's still no logs seen in the admin panel.

max.bra commented on 2017-01-10 10:11 (UTC)

hi visibilityspots, thanks for reporting but all is functional here. can you post a little more debug informations? archlinux flavour?

visibilityspots commented on 2017-01-10 09:15 (UTC)

the ttl option in the /etc/dnsmasq.d/01-pihole.conf prevents the dnsmasq.service from starting.. By removing it from the config it starts without any issue.

max.bra commented on 2017-01-06 09:35 (UTC)

@Boohbah already solved with last revision. Thanks for reporting.

commented on 2017-01-06 01:23 (UTC)

This line was being spammed in the logs. lighttpd-angel[321]: sh: line 0: cd: /etc/.pihole/: No such file or directory Solved by removing references to /etc/.pihole in /srv/http/pihole/pihole/index.php $piHoleVersion = exec('cd /etc/.pihole/ && git describe --tags --abbrev=0'); <footer>Generated <?php echo date('D g:i A, M d'); ?> by Pi-hole <?php echo $piHoleVersion; ?></footer>

max.bra commented on 2017-01-04 23:19 (UTC)

@dafero @L54 YES!! you are right. I will correct it ASAP. Thanks for reporting.

dafero commented on 2017-01-04 15:30 (UTC) (edited on 2017-01-04 15:31 (UTC) by dafero)

configuration file update instructions are out of date: L54 says: # [ -f /etc/dnsmasq.orig ] && cp /etc/pihole/configs/dnsmasq.complete /etc/dnsmasq.conf and it should be: # [ -f /etc/dnsmasq.orig ] && cp /etc/pihole/configs/dnsmasq.main /etc/dnsmasq.conf

max.bra commented on 2017-01-01 17:33 (UTC)

thanks to C0rn3j I can afford to submit a pre-package for those of you who have time to test it feedbacks to my profile mail will be appreciated.

max.bra commented on 2017-01-01 17:24 (UTC)

Update to ver. 2.10.x need manual interaction: delete /etc/dnsmasq.d/01.pihole.conf is needed before updating. From ver. 2.10.x this package will manage the files in the folder dnsmasq.d/.

max.bra commented on 2016-12-31 18:21 (UTC)

why not, will try to do a package tomorrow then. happy new year to you all!!!

C0rn3j commented on 2016-12-31 18:09 (UTC)

Thanks for supporting the package and happy New Year! I can possibly test it on my RPi when you'll have time to work on it, email me if you'd like to ^^

max.bra commented on 2016-12-27 23:07 (UTC)

Hi C0rn3j, this package is not abandoned. I'm not able to update and test it at the moment (christmas holidays and broken pi *facepalm*) Basic functionality of pi-hole are really the same (of course) but web interface (mostly) and core are very different from current package version with lots of new features and services. It is not enough to change version number and MD5 sums. Please, be patient... ... and pray for my pi :-)

C0rn3j commented on 2016-12-27 22:13 (UTC)

Tried updating it myself by changing these lines pkgver=2.10.2 _wwwpkgver=2.2 md5sums=('eb6f32a75bdc39825119b017b79b2746' 'f10faa6542a48c78348ab6000be5db31' But it fails like this, any ideas? ==> Sed error: sudoers setup 1 ==> ERROR: A failure occurred in prepare().

ThecaTTony commented on 2016-11-25 20:20 (UTC)

Sorry, it's not a complaint, it's to improve the package. I appreciate the work you do here. I think the best way to to solve the warnings of the .install is to hide them. This are the modifications i made, let me know what you think about:

max.bra commented on 2016-11-25 09:21 (UTC) (edited on 2016-11-25 09:34 (UTC) by max.bra)

@ThecaTTony Anything else? > * Also pi-hole-server.install is wrong because assumes that dnsmasq.d and 01-pihole.conf exist indeed, it's that assumes that. it's a little recursive problem as the package leave to the user final configuration of dnsmasq. Do you prefere i forced replace your config files?

ThecaTTony commented on 2016-11-22 21:03 (UTC) (edited on 2016-11-23 17:53 (UTC) by ThecaTTony)

Hi, thanks for packaging. Please follow Web Application Package Guidelines at: * webapp should be on /usr/share/webapps/pihole * dont force libmariadbclient lighttpd and php-cgi as mandatory dependencies. Pi-Hole can run without the need of the admin interface and/or with nginx/php-fpm as webserver. EDIT (11/23/2016): * Also pi-hole-server.install is wrong because assumes that dnsmasq.d and 01-pihole.conf exist: ==> Neutrino emissions detected... cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory /usr/bin/pihole: line 126: /etc/dnsmasq.d/01-pihole.conf: No such file or directory cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory /usr/bin/pihole: line 126: /etc/dnsmasq.d/01-pihole.conf: No such file or directory

commented on 2016-11-18 07:09 (UTC)

Hey, great packaging! This is Dan, one of the Pi-hole developers. Send me an email at (my key is 712CEDA). We'd like to see about collaborating a bit and helping out if we can.

max.bra commented on 2016-11-03 07:05 (UTC)

@tdlk i agree. seems that pihole project, sooner or later, want to implement log history management. secretly hoping for sqlite.

tdkl commented on 2016-11-02 16:45 (UTC) (edited on 2016-11-02 16:47 (UTC) by tdkl)

> hi tdkl, > upstream added mod_auth to lighttpd modules, probably for future purpose. > mod_auth need libmariadbclient. Ah yeah, that made sense. I've skipped the inclusion of the mod_auth in my conf. Also I hope they won't make something stupid like making it mandatory for pi-hole to use a SQL database, because that running a RDBMS would kill its usage on low memory ARM devices.

max.bra commented on 2016-11-02 08:44 (UTC)

The bill is on me. :-D

slip commented on 2016-11-02 08:43 (UTC)

Nevermind, that last one was an error on my part. IT'S ALL FIXED! Let's go get a pint!

max.bra commented on 2016-11-02 08:42 (UTC)

ok, that is admin side. last two are (should) for your empty log and are not critical first is. but, more important, are your ads piholed again?

slip commented on 2016-11-02 08:37 (UTC)

Eh, I need to be able to figure this stuff out better, so I can't blame you. Although my setup still isn't working :/ I get this timeout for the web interface ni journalctl Nov 02 09:33:40 pihole lighttpd-angel[269]: 2016-11-02 09:33:39: (mod_fastcgi.c.2534) FastCGI-stderr: PHP Fatal error: Maximum execution time of 30 seconds exceeded in /srv/http/pihole/admin/data.php on line 26 Nov 02 09:33:40 pihole lighttpd-angel[269]: 2016-11-02 09:33:39: (mod_fastcgi.c.2534) FastCGI-stderr: PHP Warning: min(): Array must contain at least one element in /srv/http/pihole/admin/data.php on line 257 Nov 02 09:33:40 pihole lighttpd-angel[269]: 2016-11-02 09:33:39: (mod_fastcgi.c.2534) FastCGI-stderr: PHP Warning: max(): Array must contain at least one element in /srv/http/pihole/admin/data.php on line 256

max.bra commented on 2016-11-02 08:09 (UTC) (edited on 2016-11-02 08:25 (UTC) by max.bra)

@10479 yes my fault! there's something missing in dnsmasq conf updating... done! with new 2.9.5 dnsmasq config files i lost the include line! my pi manage dhcp part too and using a conf of mine i missed completely the serious mistake. sorry for wasting your time.

max.bra commented on 2016-11-02 07:57 (UTC) (edited on 2016-11-02 07:59 (UTC) by max.bra)

journalctl should say something... pihole -c output? forgot: yes, all is working here

slip commented on 2016-11-02 07:51 (UTC)

Yeah. I've been using this package since you first released it and have always been well off with the instructions. Both dnsmasq.conf and 01-pihole.conf are in their place on a literally fresh and fully up to date install of arch. I don't know. I give up. I've wasted entirely too much time on this. I've installed it about 20 times today including on a fresh arch install. Clearly something is wrong because the margin for error is incredibly slim on a fresh install with nothing else to interfere. But ads aren't being blocked, logs aren't being created and the interface reflects it.

max.bra commented on 2016-11-02 07:16 (UTC)

have you followed the configuration instructions? lighttpd.conf is new dnsmasq conf files, now, are two: dnsmasq.main -> /etc/dnsmasq.conf dnsmasq.include -> /etc/dnsmasq.d/01-pihole.conf

slip commented on 2016-11-02 05:01 (UTC) (edited on 2016-11-02 07:03 (UTC) by slip)

Edit - I've even nuked the entire pi and started from scratch since it's only used for pihole. Still no pi.hole/admin resolution and the web interface still hangs and does nothing. Maybe the web interface is an upstream issue as they've had that issue in the past. Even further analysis shows that no logs are being generated.

max.bra commented on 2016-11-01 21:23 (UTC)

@10479 it's a little recursive problem. upgrade follow post upgrade instruction (creating /etc/dnsmasq.d/01-pihole.conf) reinstall again or run pihole -g

slip commented on 2016-11-01 20:01 (UTC) (edited on 2016-11-01 20:03 (UTC) by slip)

Wow, fast response time!! I ran the latest update, but I'm still not getting pi.hole/admin to work. Are you? When I install, it complains about this, but to be fair, that was there after the last update as well. cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory /usr/bin/pihole: line 126: /etc/dnsmasq.d/01-pihole.conf: No such file or directory Which is tied to: #add addn-host= to dnsmasq echo "addn-hosts=/etc/pihole/gravity.list" >> /etc/dnsmasq.d/01-pihole.conf It looks like this whole 01-pihole.conf is for that new enable/disable feature, but maybe it's causing more issues than it's trying to solve for us? Edit: And when I update pihole with pihole -g, I get this: ::: Refresh lists in dnsmasq... cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory cat: /etc/dnsmasq.d/01-pihole.conf: No such file or directory ::: No hosts file linked to dnsmasq, adding it in enabled state /usr/bin/pihole: line 126: /etc/dnsmasq.d/01-pihole.conf: No such file or directory

max.bra commented on 2016-11-01 19:29 (UTC) (edited on 2016-11-01 19:29 (UTC) by max.bra)

@10479 good point! thanks for reporting. completely missed that! done following official way. edit: typo

slip commented on 2016-11-01 18:44 (UTC)

One of the tweaks listed for the latest release is: [TWEAK] - Remove pi.hole and <hostname> from gravity.list put them as address= in 01-pihole.conf #849 I'm assuming that is what's causing the pi.hole/admin interface to stop working in the AUR release as I don't see that file, nor pi.hole referenced in the PKGBUILD. Not a huge deal, but the name resolution was nice, because just <hostname>/admin doesn't allow changes to be made in the web interface. Should pi.hole just be added back into the gravity.list?

max.bra commented on 2016-11-01 16:44 (UTC)

hi tdkl, upstream added mod_auth to lighttpd modules, probably for future purpose. mod_auth need libmariadbclient.

tdkl commented on 2016-11-01 15:52 (UTC)

Why is libmariadbclient set as dep now ?

max.bra commented on 2016-10-14 15:21 (UTC)

hi peoro, i think you have posted in wrong place: where can you find a pod2man command in this PKGBUILD?

peoro commented on 2016-10-14 15:18 (UTC) (edited on 2016-10-14 15:18 (UTC) by peoro)

This package should have a build dependency to whatever provides `pod2man`: ==> Starting build()... PKGBUILD: line 16: pod2man: command not found ==> ERROR: A failure occurred in build(). Aborting...

max.bra commented on 2016-10-05 11:51 (UTC)

sorry my fault. updated without version bump.

TimLovesTech commented on 2016-10-05 08:52 (UTC)

The file "" has been changed in the latest snapshot, and the md5 was not updated so it fails validity check.

nikke commented on 2016-08-30 08:39 (UTC)

It's black for me aswell when I open the logo locally with a image viewer, tested several. But when I open logo.svg with palemoon or chrome it's shown. My fix is for the logo showing as broken in the web interface.

max.bra commented on 2016-08-30 07:48 (UTC)

hi nikke, thanks a lot for your advice. but this is inside web interface: i was asking for svg appearance outside it. > wget

nikke commented on 2016-08-30 07:38 (UTC)

same here, added ".svg" => "image/svg+xml", to lighttpd.conf. restarted lighttpd and logo is shown OK.

max.bra commented on 2016-08-24 12:38 (UTC)

@all hi all, i need a little team test. Seems there is something wrong with new pi-hole php admin package. The new logo (SVG format) is... black! Or at least it's black for me BUT only if i open it locally, inside browsers it is right. can you please execute this below and let me know what you see/obtain? wget

slip commented on 2016-07-12 00:22 (UTC)

Came down to the fact that I was using the hostname for the web interface. Someone over on reddit just posted about the same issue and said using the IP address instead of host name works, and that's true. So it's an issue on the end of pi hole.

slip commented on 2016-07-07 14:58 (UTC)

@eleete0712 Thanks for the confirmation. I'll give a manual install a shot and see what happens.

eleete0712 commented on 2016-07-06 06:53 (UTC)

@10479 Having no problems adding and removing entries to the whitelist / blacklist via the web interface. So it is no general issue. Running on Arch Linux Arm (ODROID-C2), installed manually not using yaourt.

slip commented on 2016-07-05 01:03 (UTC)

Sorry, been busy. So I just did a fresh install of Arch ARM (pi 2) and pi-hole-server. I still can't add or remove from the web interface. Exact steps after install were adding a user account, yaourt and then pi-hole-server and setting it up according to the instructions. All works well except the web interface for white and blacklists. It would be good if someone else here confirmed if they had the issue or not.

max.bra commented on 2016-06-28 21:08 (UTC)

> It's acting as if it doesn't have permissions. yes it is, in lighttpd session.save_path parameter path precisely > Is there an error log for pihole? lighttpd and dnsmasq are logging into systemd journal as usual

slip commented on 2016-06-28 14:15 (UTC)

I haven't had a chance to reinstall yet. I might be able to today. Is there an error log for pihole? I looked a bit but didn't see one. It's acting as if it doesn't have permissions.

max.bra commented on 2016-06-28 06:38 (UTC)

since changes of the new version are limited to adlists.default we can also upgrade without any test. still renovation, be patient. @10479 hi 10479. how's the web interface, you solved? @all anyone else with [white,black]list web interface problem?

slip commented on 2016-06-28 00:32 (UTC)

Just a heads up, I just updated pihole and it reverted back to a version of an ad list that is no longer function as http, and now requires https. should now be https. after changing it to https manually, you can 'rm /etc/pihole/list.*' and then 'pihole -g' to regain it's functionality.

max.bra commented on 2016-06-23 07:41 (UTC)

Hi promofaux! Sorry for late replay, home renovarion nightmare here!!!! > would you mind lending your Arch experience to the thread? Sure i will, ASAP in maximum 1-2 days.

promofaux commented on 2016-06-22 13:00 (UTC)

Hi max.bra, Promofaux here from pi-hole. Just had a user mention on reddit that this package doesn't give a choice to select IPv6 on install, I've not actually used this (total arch noob here), but a couple of things I suggested to him do not seem to work.. would you mind lending your Arch experience to the thread?

max.bra commented on 2016-06-05 14:35 (UTC)

> I'm definitely not a PHP guru same i've ever seen an "open_basedir" of a single file... PHP docs: " the specified directory-tree, including the file itself.". debatable for me... anyway, updating to this easier and maintainable solution.

napgravy commented on 2016-06-05 13:54 (UTC) (edited on 2016-06-05 13:57 (UTC) by napgravy)

I'm definitely not a PHP guru, but the PHP documentation suggests a single file can be specified. I've added "/etc/hosts" here and it seems okay, but to test it out, I modified the Pi-Hole data.php file to add another "file_exists" check on a different file in /etc/. ie. $hostname = file_exists("/etc/hostname") ? file("/etc/hostname") : array(); I started the admin page in my browser and sure enough, journal errors were generated indicating "/etc/hostname" isn't within the allowed path(s). Unless I'm missing something, it appears you don't have to expose all of "/etc" -- just "/etc/hosts". :) EDIT -- just noticed you solved it another way in your recent changes. :)

max.bra commented on 2016-06-04 20:03 (UTC)

hi napgravy, thanks for reporting. you are right: [ $hosts = file_exists("/etc/hosts") ? file("/etc/hosts") : array(); ] in data.php line 4, committed on 16 Apr (02f6517b55c59b76f928ed52c476109e2d4c2c4a) is creating the issue you submitted... not as easy as it looks: to allow /etc/hosts we need to open_basedir ALL /etc directory, and is not really a good idea although presumably none of our "pihole" is publicly exposed to the network. give me some time to see what can be done without distorting the original project. suggestions are welcome.

napgravy commented on 2016-06-04 15:59 (UTC)

Hello! /srv/http/pihole/admin/data.php attempts a "file-exists" on "/etc/hosts" (line 4) When you are viewing the pi-hole admin page in your browser, this generates many errors in the journal because "/etc/hosts" is not in the open_basedir path. Easy enough to fix by adding "/etc/hosts" to the open_basedir path in lighttpd.conf. :) ...something to consider for a future package update.

max.bra commented on 2016-04-21 05:38 (UTC)

hi atraii, thanks for reporting. It happened with commits of 15 and 16 January to make whitelist and blacklist independent. I have not seen them... :-(

Atraii commented on 2016-04-21 00:20 (UTC)

The helper scripts are using the debian service commands (/usr/bin/ and don't work correctly to restart the systemd service. Just a heads up while you're de-debian'izing the installer.

max.bra commented on 2016-04-13 06:56 (UTC)

hi zer0t3ch, thanks for reporting. done.

zer0t3ch commented on 2016-04-13 00:03 (UTC)

I also just submitted an issue about this on their GitHub tracker.

zer0t3ch commented on 2016-04-12 23:32 (UTC)

Should possibly be modified in this package for us? Line 22 is using dpkg-query.

max.bra commented on 2016-04-11 14:04 (UTC)

hi zerojay! > > > Probably should bump the package due to this major bug in 2.6.1. sure i should!! thanks for reporting. > Also, I cannot seem to find, added in 2.6: this package, as description said, is an adaptation of original pi-hole distro for archlinux. original pihole distro support raspbian. is merely a wrapper for other commands already present in this package. and also, manage commands absolutely not supported (uninstall for example) and absolutely not to be executed. for now, is unnecessary and dangerous.

zerojay commented on 2016-04-11 13:49 (UTC)

Also, I cannot seem to find, added in 2.6:

zerojay commented on 2016-04-11 12:27 (UTC) Probably should bump the package due to this major bug in 2.6.1.

max.bra commented on 2016-03-08 11:11 (UTC)

@tdkl sob, basic lighttpd configuration file is not really adaptable to additional include

tdkl commented on 2016-03-08 10:58 (UTC)

@max.bra: yeah dokuwiki rocks. No idea what to do about the config, because if someone uses a custom config with many websites, it's going to take some manual work anyway. Unless you change the pi-hole to a location that is already listed under default open_basedir config.

max.bra commented on 2016-03-08 07:47 (UTC)

@tdkl mmm, do you use lighttpd for dokuwiki too? do you think is better to mimic dnsmasq config type (complete/include) for lighttpd too?

tdkl commented on 2016-03-08 07:17 (UTC) (edited on 2016-03-08 07:18 (UTC) by tdkl)

Forgot to test dokuwiki with the open_basedir conf in lighttpd.conf, it stops working. Dokuwiki is in /dokuwiki. But no big deal, I've changed php.ini again, so now both are working again, just a heads up.

max.bra commented on 2016-03-02 15:17 (UTC) (edited on 2016-03-02 15:18 (UTC) by max.bra)

@bolorino @tdkl very well then. thanks to you all. see you beyond the event horizon. :-D edit: typo

bolorino commented on 2016-03-01 21:48 (UTC)

@max.bra Fine. It was my fault not catching the upgraded config files message. I'd restarted the services but using the old config files. It seems to be running fine now. Thanks again!

max.bra commented on 2016-03-01 17:03 (UTC) (edited on 2016-03-01 17:24 (UTC) by max.bra)

@bolorino > It is possible that the data is being recorded at /run/log/pihole.log and the panel is looking for it at pihole dir under /run/log? YES, FOR SURE as i stated in update message: ==> ver. 2.5.3-3: dnsmasq and lighttpd conf files are changed, please repeat the installation steps you are using second-last dnsmasq and lighttpd config files and/or you have not restarted them repeat installation steps please and let's see what happens. p.s.: old /run/log/pihole.log will be purged by next reboot as it's in a tmpfs path

bolorino commented on 2016-03-01 15:37 (UTC) (edited on 2016-03-01 15:39 (UTC) by bolorino)

I've found the problem with the dashboard no getting the data. It seems we have /run/log/pihole.log but also /run/log/pihole/pihole.log It is possible that the data is being recorded at /run/log/pihole.log and the panel is looking for it at pihole dir under /run/log? Edit: typo

bolorino commented on 2016-03-01 15:14 (UTC)

@max.bra You are right. I have no mysql installed. Strange is that it was working without it. I also have some fully functional PHP services through php-fpm under nginx. I get the mysql error with lighthttp but no with nginx. Will check further. Thank you.

max.bra commented on 2016-03-01 14:49 (UTC)

assuming you have followed pi-hole configuration steps, there is something wrong with your php setup

bolorino commented on 2016-03-01 14:44 (UTC)

@max.bra: No luck with the latest update. The panel isn't working on nginx, neither lighthttp. On nginx I get the empty values and widgets. I've stopped nginx and launched lighthttp and get: > lighttpd-angel[26035]: PHP Warning: PHP Startup: Unable to load dynamic library '/usr/lib/php/modules/' - /usr/lib/php/modules/ cannot open shared object file: No such file or directory in Unknown on line 0 Maybe is better to leave it as intended in pihole setup? One always can fine tune their own web server setup to get it working. After all it is something beyond the scope of the project, I think. :-)

tdkl commented on 2016-03-01 12:37 (UTC)

@max.bra : tested it and it works, thanks :]

max.bra commented on 2016-03-01 11:51 (UTC) (edited on 2016-03-01 11:51 (UTC) by max.bra)

-3 update This should resolve the "issue" of lighttpd php. Your open_basedir can be reverted without taking into account pi-hole need. edit: still available for a possible nginx integration

max.bra commented on 2016-03-01 10:08 (UTC)

@tdkl @bolorino yep yep. for me, i have no other web servers on my pi and my open_basedir is not set allowing all (following pihole setup) can be improved...

tdkl commented on 2016-03-01 09:49 (UTC)

Just to chime in: I'm using the default lighttpd server and the admin panel didn't work until I added :/etc/pihole/:/run/log/ to open_basedir.

max.bra commented on 2016-03-01 09:33 (UTC)

@bolorino reverting open_basedir can you try this? p.s.: to be refined and improved

bolorino commented on 2016-02-29 23:42 (UTC)

@max.bra: The problem was the open_basedir directive of PHP in /etc/php/php.ini > mar 01 00:05:44 helicon nginx[1244]: PHP message: PHP Warning: file(/etc/pihole/gravity.list): failed to open stream: Operation not permitted in /srv/http/pihole/admin/data.php on line 144 The panel works fine after adding /etc/pihole/ and /run/log to open_basedir. Using other “web apps” that store the configs needed in /etc/webapps/{app_name} (dokuwiki for example), but I don't know if this is good practise. I have no experience with AUR packaging. To me it seems better having a single, unified path outside the webserver (/etc/webapps/ for instance) if needed, or even better, having none. Anyway if the files needed for the panel to work are outside of the web server path (/srv/http for instance), you are going to need to give PHP access to them. Sorry for any typos or mistakes. English is not my language. And thanks again :-)

max.bra commented on 2016-02-29 22:21 (UTC)

i never used nginx but seems not too difficult to integrate pi-hole in it. if you agree, i'm available to insert an nginx include conf file, if feasible. keep me posted.

bolorino commented on 2016-02-29 21:28 (UTC) (edited on 2016-02-29 21:32 (UTC) by bolorino)

Well. More likely to be me using nginx along with lighttpd, which should not be neccesary in my setup. Will check. Edit one word.

bolorino commented on 2016-02-29 21:23 (UTC)

@max.bra: Wow! Fast. Thank you :-) There are some problems with the panel not showing the information and empty widgets. It seems to be related to open_basedir restrictions. The paths are: /etc/pihole/gravity.list /run/log/pihole.log I will check my PHP config and the lighthttp one.

max.bra commented on 2016-02-29 21:05 (UTC)

aaargh! i was really fall behind with php interface!! thanks for reporting. updated: let me know if something is not working.

bolorino commented on 2016-02-29 19:02 (UTC)

There is a nice update for the Admin panel (AdminLTE), 1.0.1: Would it be possible to include it with the pi-hole v2.5.3? Thanks!

max.bra commented on 2016-02-18 12:49 (UTC)

> But I guess it's better to leave LC_TIME to en_US.UTF-8 to avoid problems shown today with the capital letter for the month. sure! > Regarding the subliminal package, it's a python package that requires some other python packages, which some failed to compile because I didn't have the proper locale set at the time. if they are so squeamish you can apply the LC_*DUMMY*=C %command% trick to configure and or make commands of their PKGBUILDs

tdkl commented on 2016-02-18 12:41 (UTC)

@max.bra: good news, the reboot seems to have worked, now also the admin site displays the numbers. Guess the good ol' reboot saves the day. Thanks again ! Regarding the subliminal package, it's a python package that requires some other python packages, which some failed to compile because I didn't have the proper locale set at the time. So when I looked into setting the locale, I changed them mostly to my country UTF-8 locale, only LC_COLLATE was left at C. But I guess it's better to leave LC_TIME to en_US.UTF-8 to avoid problems shown today with the capital letter for the month.

max.bra commented on 2016-02-18 12:30 (UTC)

... continued maybe you should force locale for that subliminal package... standard wins!!

max.bra commented on 2016-02-18 12:27 (UTC)

@tdkl for system wide locale changes you should apply a reboot problem is not your locale, but dnsmasq that seems not to honor it let's see what happens when you apply all your relevant locales to a standard en_US.UTF-8 and reboot your pogo...

tdkl commented on 2016-02-18 12:11 (UTC)

@max.bra: Interestingly if I set LC_TIME to default LANG which is en_US.UTF-8: [tdkl@pogo system]$ locale LANG=en_US.UTF-8 LC_CTYPE=en_US.UTF-8 LC_TIME="en_US.UTF-8" LC_COLLATE=C LC_ALL= The default still works: [tdkl@pogo system]$ /usr/bin/ -j {"domains_being_blocked":"145791","dns_queries_today":"4971","ads_blocked_today":"340","ads_percentage_today":"6.83"} But the admin website displays 0 again. If I add your change to display date with C in, the website works. This is really odd because the JSON output clearly gives out the numbers, but website doesn't display them.

tdkl commented on 2016-02-18 11:47 (UTC)

@max.bra: sweet, this displays the date with capital letter and DNS queries and blocked % is displayed again :] Thanks a lot. I don't imagine you'll set this to for everyone though. Do you suggest I change the LC_TIME locale back to C, so I don't have to edit the at each update ?

max.bra commented on 2016-02-18 11:38 (UTC)

done. @tdkl can you try this? .... piLog="/run/log/pihole.log" gravity="/etc/pihole/gravity.list" today=$(LC_TIME=C date "+%b %e") ....

tdkl commented on 2016-02-18 11:01 (UTC)

@max.bra : it seems the main issue is that in my locale: [tdkl@pogo ~]$ date "+%b %e" feb 18 Gives out month name started with a non-capital letter. While the ones in pihole.log are started with the capital letter for the short name of the month : Feb 18 11:59:43 dnsmasq[239]: cached is So I'd have to change the date output somehow to put out the capital letter to avoid changing the date locale. Linux ¯\_(ツ)_/¯

tdkl commented on 2016-02-18 10:56 (UTC)

@max.bra : bingo, I think you've found the issue. [tdkl@pogo log]$ cat "/run/log/pihole.log" | grep "$today" | awk '/query/ {print $6}' | wc -l Although I derped and forgot to change the variable for $today to an actual value when testing in terminal, just as for pihole.log, this didn't result in 0 value. But if I use the date on my machine : [tdkl@pogo ~]$ date "+%b %e" feb 18 Which is a result of my locale. I changed locales a bit when setting up subliminal package, so I guess I need to change the date one back to UTF8.en. Thanks for the time and the tip :]

max.bra commented on 2016-02-18 10:49 (UTC)

@tdkl (second-last) given this: > [tdkl@pogo log]$ cat "/run/log/pihole.log" | grep "$today" | awk '/query/ {print $6}' | wc -l > 4000 it is definitely a bc rounding problem, but why? your locale/locale conf?

max.bra commented on 2016-02-18 10:41 (UTC)

@tdkl (last one) I saw an excessive filtration and was considering removing the latest updates of the list; given the reporting and the commits I do it as soon as possible.

tdkl commented on 2016-02-18 10:32 (UTC)

@max.bra: Wanted to report this separately, there were 2 commits who remove 2 default lists. One of them is empty, but one of them breaks sites. I was wondering why was suddenly down for couple days, but didn't thought it could be the pihole blocking it. After removing the site, I could access it again. So IMHO those two commits should be pushed as an update. Empty list: Blocks extra content:

tdkl commented on 2016-02-18 10:31 (UTC)

@max.bra : Nope, queries are logged fine to /run/log/pihole.log [tdkl@pogo log]$ tail -n5 pihole.log Feb 18 11:12:00 dnsmasq[239]: query[A] from Feb 18 11:12:00 dnsmasq[239]: /etc/pihole/gravity.list is Feb 18 11:12:02 dnsmasq[239]: query[A] from Feb 18 11:12:02 dnsmasq[239]: cached is Feb 18 11:12:02 dnsmasq[239]: cached is Also nothing changed regarding dnsmasq.conf. I use an additional .conf for pihole, so I've added this at the end of dnsmasq.conf: # Include another lot of configuration options. conf-dir=/etc/dnsmasq.d Then have a special pihole.conf there: [tdkl@pogo log]$ cat /etc/dnsmasq.d/pihole.conf addn-hosts=/etc/pihole/gravity.list cache-size=10000 log-queries log-facility=/run/log/pihole.log log-async So no clue. I've checked the and piLog var is set to /run/log/pihole.log. Running the part of the function CalcQueriesToday() manually gives out proper result: [tdkl@pogo log]$ cat "/run/log/pihole.log" | grep "$today" | awk '/query/ {print $6}' | wc -l 4000 Function CalcPercentBlockedToday() gives out 0, because apparently $queriesToday calculated in upper function is 0. But running the gives out 0: [tdkl@pogo ~]$ /usr/bin/ -j {"domains_being_blocked":"145787","dns_queries_today":"0","ads_blocked_today":"240","ads_percentage_today":"0"} Shouldn't also be a read permission issue for pihole.log: -rw-r--r-- 1 dnsmasq root 1061770 feb 18 11:25 pihole.log

max.bra commented on 2016-02-17 21:41 (UTC)

weird... does your dnsmasq log queries to /run/log/pihole.log? maybe an unintentional overwriting/modification of main .conf or include part?

tdkl commented on 2016-02-17 15:01 (UTC)

@max.bra : even after one day, those numbers are at zero, while the ads blocked indicator increments. Not a big deal though, will see how it behaves in the future.

max.bra commented on 2016-02-16 21:30 (UTC)

@tdkl: no, admin website regular here... maybe your log was truncated at that precise moment?

tdkl commented on 2016-02-16 20:07 (UTC)

@max.bra: do you also have 0 dns queries and 0 traffic % displayed after the 2.5.2 update in the Pi-hole admin website ? Seems they played with the a bit :

tdkl commented on 2016-02-15 23:08 (UTC)

You're setting your routers primary DNS to one on internal network? Jeez.. How about you check up on Arch wiki about network configuration first? Because this isn't an issue with this package.

Matty_r commented on 2016-02-09 21:54 (UTC)

It is only unable to ping the router as soon as I change the DNS server on the router to point to the pihole machine.

max.bra commented on 2016-02-09 13:55 (UTC)

with your configuration pihole machine MUST ping your router!! hardware related?

max.bra commented on 2016-02-09 11:28 (UTC)

pihole machine can ping

Matty_r commented on 2016-02-09 11:24 (UTC) (edited on 2016-02-09 11:31 (UTC) by Matty_r)

Yes, that's a typo. Edit: It doesn't appear to be able to ping, nor my router at However it can ping other computers on the network.

max.bra commented on 2016-02-09 10:29 (UTC) (edited on 2016-02-09 10:32 (UTC) by max.bra)

> **** > /etc/resolv.conf = > nameserver <----- this is a typo yes? > nameserver > **** pihole machine can ping

Matty_r commented on 2016-02-09 10:23 (UTC)

@max.bra: Ok, so i'm still having the same issue. Router was set to pri **** /etc/resolv.conf = nameserver nameserver **** Static IP is done through systemd.networkd **** /etc/systemd/network/ = [Network] DHCP=no DNS= DNS= [Address] Address= [Route] Gateway= ******** /etc/dnsmasq.conf - was replaced by supplied version /etc/lighttpd/lighttpd.conf - was replaced by supplied version troubleshooting # dig connection timed out; no servers could be reached # nslookup connection timed out; no servers could be reached

Matty_r commented on 2016-02-09 07:51 (UTC)

@max.bra: A couple of differences of what I had: Router: DNS pri - sec piHole: DHCP enabled but was reserved. It may have been dhcpcd was messing with the config, may need to make sure it's disabled. Give me some time and i'll get back to you with the results. Thanks for your time.

max.bra commented on 2016-02-09 07:43 (UTC) (edited on 2016-02-09 07:44 (UTC) by max.bra)

:: Sure i will say the obvious and you will forgive me for this :: Router: Static LAN already configured. DHCP server enabled. Relevant DHCP options: Lease - DNS pri - sec Clients: DHCP client enabled Pi-Hole: DHCP client disabled. Static NIC configuration: IP addr. DNS pri - sec dnsmasq configured via dnsmasq.complete. enabled and started (see configuration) ---- have we missed something or the problem is more subtle? edit: typo

Matty_r commented on 2016-02-09 07:25 (UTC)

@max.bra: Router: (DNS relay & DHCP server) piHole: Various other clients are from .1 to .100. I configured my Router to point to piHole didn't appear to be retrieving DNS entries and then I added to the router as a secondary DNS server. All other clients were able to connect at this point. However piHole still wasn't working. I connected to piHole and it wasn't able to ping nor, I then attempted to use pacman but it wouldn't connect either. I could however ping

max.bra commented on 2016-02-09 07:00 (UTC)

hi Matty_r, can i have your lan topology please?

Matty_r commented on 2016-02-09 03:43 (UTC)

I wasn't able to get this working. Is there anything else I need to do a part from what is in the configuration file, on a clean build? I followed all of the instructions but it wasn't able to get internet access at all after enabling everything. Do I need to disable dhcp for this to work? I have a IP address reservation set up on my DHCP server.

tdkl commented on 2016-02-05 09:15 (UTC)

Ah I see now. Take a look of what modules it loads and if the kernel is somehow modified, can't see a reason why this couldn't be applied on regular Arch install. But many developers nowadays just ship their own "distro" instead.

slackline commented on 2016-02-05 08:47 (UTC)

Ah, thanks for the pointer on pacman @tdkl I'll look into using whether Archphile can have this resolved, developer seems active, but will also consider plain archlinuxarm, will need to read around about how to get I2S DAC card working though (it was the convenience of having it supported out of the box in Archphile that appealed). Thanks for your help.

tdkl commented on 2016-02-04 10:50 (UTC) (edited on 2016-02-04 10:50 (UTC) by tdkl)

@slackline: you can see information about not installed packages with : # pacman -Sii dnsmasq <snip> Depends On : glibc libdbus gmp nettle <snip> I use the Arch Linux ARM as well, on a armv5 device (, dnsmasq installs fine. Don't know Archphile though, but it might be better if you move from it if you can to avoid possible issues.

slackline commented on 2016-02-04 10:39 (UTC)

> no, nettle is an official dep of dnsmasq > Ah, this may be why its not pulled in then, Archphile isn't x86_64 its ARM and is built on so perhaps the PKGBUILD for dnsmasq doesn't have nettle as a dependency. The package is listed when searching for 'dnsmasq' at but this doesn't seem to allow viewing of the PKGBUILD itself. Great work on the switch to systemd-timer too, thanks.

tdkl commented on 2016-02-03 14:55 (UTC)

Sweet, thanks again and same to you :]

max.bra commented on 2016-02-03 14:51 (UTC)

done! > I checked with "systemctl list-timers" after updating the package and they weren't active. yes they do, they weren't started as you rightly said > In my experience when creating a .service and .timer I always enabled/started them manually. indeed it is!! the package enables statically the timers and the .install start them. thanks again for your comments. Good weekly "::: Neutrino emissions detection"... ( quote) :-D

tdkl commented on 2016-02-03 14:37 (UTC)

I checked with "systemctl list-timers" after updating the package and they weren't active. I don't know if that gets done automatically after a while though, if you link them. In my experience when creating a .service and .timer I always enabled/started them manually.

max.bra commented on 2016-02-03 14:27 (UTC)

timers are statically enabled by the package, see ln command in PKGBUILD package() section. at least they have to be started... reboot it's not an option... it's a shame to get it done by the user, the installer can do it right. adjusting...

tdkl commented on 2016-02-03 14:04 (UTC)

Excellent, that was fast, thank you very much for bringing pi-hole to Arch. The timers still aren't activated automatically though, so the configuration file should include instructions how to activate/deactivate them (if it's not done by installing the package), like you did for dnsmasq.service and lighttpd.service. # systemctl enable pi-hole-gravity.timer # systemctl start pi-hole-gravity.timer # systemctl enable pi-hole-logtruncate.timer # systemctl start pi-hole-logtruncate.timer And stop/disable for uninstall.

max.bra commented on 2016-02-03 13:24 (UTC)

good observation!! done updating, thanks.

tdkl commented on 2016-02-03 09:17 (UTC)

I'd like to point out that in the configuration manual the notion of starting and enabling cronie/fcron/cron service is missing, since the default cron for Arch is systemd-timer. It would also be "cleaner" if systemd-timer would be used to run the daily and weekly script and the dependency on *cron would be removed.

max.bra commented on 2016-02-02 16:06 (UTC)

> Do you mean in the pi-hole-server PKGBUILD? no, nettle is an official dep of dnsmasq

slackline commented on 2016-02-02 15:31 (UTC)

Hi max.bra, Yes, typical that the Archphile site is down when I'm trying to do a fresh install. > nettle is a deps in PKGBUILD Do you mean in the pi-hole-server PKGBUILD? If so I've just had a look and being unfamiliar with the structure of PKGBUILD's think the relevant line omits nettle... depends=('cron' 'dnsmasq' 'lighttpd' 'php-cgi' 'bc' 'figlet') (nor is it listed above under dependencies which match the above line in PKGBUILD). Or have I misunderstood and its a dep in some other PKGBUILD (perhaps a dependency of dnsmasq that should be pulled in?). Happy to help nudge Archphile to include it as a dependency if the later is the case.

max.bra commented on 2016-01-31 23:06 (UTC)

hi slackline, thanks for your comment. Archphile site is down at the moment and i cannot see their dnsmasq package, but as you said yourself it's an arch based distro and nettle is a deps in PKGBUILD.

slackline commented on 2016-01-31 22:40 (UTC)

I think I've sorted this. Tried running dnsmasq manually and got the following... # /usr/bin/dnsmasq /usr/bin/dnsmasq: error while loading shared libraries: cannot open shared object file: No such file or directory ...installed nettle (# pacman -S nettle) and install of pi-hole-server proceeds ok... $ makepkg -sri ==> WARNING: A package has already been built, installing existing package... ==> Installing package pi-hole-server with pacman -U... loading packages... warning: pi-hole-server-2.4-2 is up to date -- reinstalling resolving dependencies... looking for conflicting packages... Packages (1) pi-hole-server-2.4-2 Total Installed Size: 0.26 MiB Net Upgrade Size: 0.00 MiB :: Proceed with installation? [Y/n] y (1/1) checking keys in keyring [#######################################################################################] 100% (1/1) checking package integrity [#######################################################################################] 100% (1/1) loading package files [#######################################################################################] 100% (1/1) checking for file conflicts [#######################################################################################] 100% (1/1) checking available disk space [#######################################################################################] 100% (1/1) reinstalling pi-hole-server [#######################################################################################] 100% ==> please read configuration instructions at /usr/share/doc/pihole/configuration ==> lighttpd config file is changed. please uninstall and reinstall following configuration steps. ==> first install/update run... $

slackline commented on 2016-01-31 18:33 (UTC)

First thanks for making this package, really useful and saves going through a manual pi-hole install (which I've done a couple of times, not impossible, but this makes it cleaner to install and maintain). I just tried installing this and the md5sum for fails (presumably because has been updated). I manually corrected this in the local PKGBUILD and install proceeds but then installation fails with... ==> please read configuration instructions at /usr/share/doc/pihole/configuration ==> lighttpd config file is changed. please uninstall and reinstall following configuration steps. ==> first install/update run... Job for dnsmasq.service failed. See "systemctl status dnsmasq.service" and "journalctl -xe" for details. I read through /usr/share/doc/pihole/configuration and manually performed the steps for first install of both dnsmasq and the general lighttpd ones but get the same result when restarting dnsmasq... # systemctl reenable dnsmasq.service Created symlink from /etc/systemd/system/ to /usr/lib/systemd/system/dnsmasq.service... # systemctl restart dnsmasq.service Job for dnsmasq.service failed. See "systemctl status dnsmasq.service" and "journalctl -xe" for details. # systemctl status dnsmasq.service �� dnsmasq.service - A lightweight DHCP and caching DNS server Loaded: loaded (/usr/lib/systemd/system/dnsmasq.service; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Sun 2016-01-31 19:25:08 EET; 12s ago Docs: man:dnsmasq(8) Process: 3253 ExecStartPre=/usr/bin/dnsmasq --test (code=exited, status=127) My understanding is that bash returns exit status of 127 when it can't find the command so I've checked that dnsmasq is where its expected to be... ls -l /usr/bin/dnsmasq -rwxr-xr-x 1 root root 296512 Aug 3 20:37 /usr/bin/dnsmasq ...and it appears to be there and the install commands at least are being executed as root (not sure if this carries through when using systemctl (fairly new to systemd, used Gentoo for years on my systems and have stuck with openrc so this is my first real encounter with systemd). Hope its ok to leave these comments here, was unsure whether to file a bug (happy to do so if thats easier/neater for package maintenance but since it might be me or the slightly tweaked system at fault* I thought I'd start here). * Background - I've just installed Archphile an Arch based streaming distro for Raspberry Pi's because it uses ympd ( rather than traditional web-servers for WebUI as I specifically wanted to run pi-hole on the same system and was finding the conflict of multiple web-servers (e.g. nginx used by moode and lighttpd) running on the same system a bit confusing to resolve as I was struggling to translate the lighttpd redirect to the nginx configuration.

slip commented on 2016-01-17 19:13 (UTC)

Okay, no worries. Thanks for pulling this over. It's working great and took all of 2 minutes to setup.

max.bra commented on 2016-01-17 13:59 (UTC)

hi 10479, as far as i know, in original index.php there is nothing beyond base information that you see (at least for now), it is a simple matter of html/css template.

slip commented on 2016-01-16 17:34 (UTC)

Man... that was easy. One question though. When I click on the more information for the admin page, nothing shows up. Is there anything more for setup that what's listed in your instructions?