Package Details: cfengine 3.15.2-1

Git Clone URL: https://aur.archlinux.org/cfengine.git (read-only, click to copy)
Package Base: cfengine
Description: Automated suite of programs for configuring and maintaining Unix-like computers.
Upstream URL: https://cfengine.com
Licenses: GPL3
Submitter: None
Maintainer: bidulock (Lex-2008)
Last Packager: bidulock
Votes: 20
Popularity: 0.000000
First Submitted: 2008-10-25 17:55 (UTC)
Last Updated: 2021-05-25 02:39 (UTC)

Latest Comments

bidulock commented on 2021-05-25 02:40 (UTC)

Not sure which two tests are failing or why, but pass --nocheck to pass compile and build package.

daebenji commented on 2020-11-16 15:48 (UTC)

Hello,

no worries. I'm glad that the package exists and you were able to fix it.

Thanks a lot.

Lex-2008 commented on 2020-11-16 14:27 (UTC)

Thank you for the report, daebenji! Sorry for a long time with no answer - is it even okay to say "Thank you for the report" after 4 months of silence?

Anyway, I indeed reproduced it and found that it was reported on Debian sid and fixed on master branch - in https://tracker.mender.io/browse/CFE-3401, patch is available here: https://github.com/cfengine/core/commit/6c07f44f554522212c01e34d23f8e40c62a8b363.patch

However, after applying this patch some acceptance tests are failing - I will look into it.

Quick-and-dirty workaround is to disable make check - the compilation error was in one of unit tests (and patch fixes the test, not the code). But I need to look into those acceptance tests failures before pushing my changes to git.

daebenji commented on 2020-07-10 11:43 (UTC)

Compiling is not working, error appears during make

../../libpromises/syslog_client.c: In function 'RemoteSysLog': ../../libpromises/syslog_client.c:118:31: warning: '%s' directive output may be truncated writing up to 1023 bytes into a region of size between 1004 and 1019 [-Wformat-truncation=] 118 | "<%i>%.15s %s %s[%ld]: %s", | ^~ ...... 122 | VPREFIX, | ~~~~~~~ ../../libpromises/syslog_client.c:118:17: note: directive argument in the range [-2147483648, 2147483647] 118 | "<%i>%.15s %s %s[%ld]: %s", | ^~~~~~~~~~~~~~~~~~~~~~~~~~ In file included from /usr/include/stdio.h:867, from ./../../libntech/libutils/platform.h:72, from ./../../libpromises/syslog_client.h:28, from ../../libpromises/syslog_client.c:25: /usr/include/bits/stdio2.h:67:10: note: '__builtinsnprintfchk' output 11 or more bytes (assuming 1059) into a destination of size 1024 67 | return builtinsnprintfchk (s, __n, __USE_FORTIFY_LEVEL - 1, | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 68 | __bos (__s), __fmt, __va_arg_pack ()); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ CCLD logging_test /usr/bin/ld: ../../libntech/libutils/.libs/libutils.a(logging.o):/home/service/.cache/yay/cfengine/src/cfengine-3.15.1/libntech/libutils/logging.c:35: multiple definition of `VPREFIX'; logging_test.o:(.bss+0x0): first defined here collect2: error: ld returned 1 exit status make[4]: [Makefile:1450: logging_test] Error 1 make[3]: [Makefile:2775: check-am] Error 2 make[2]: [Makefile:489: check-recursive] Error 1 make[1]: [Makefile:708: check-recursive] Error 1 make: *** [Makefile:997: check] Error 2 ==> ERROR: A failure occurred in check(). Aborting... error making: %!s(func() string=0x55e3fe465e80)

daebenji commented on 2020-07-10 11:43 (UTC)

Compiling is not working, error appears during make

../../libpromises/syslog_client.c: In function 'RemoteSysLog': ../../libpromises/syslog_client.c:118:31: warning: '%s' directive output may be truncated writing up to 1023 bytes into a region of size between 1004 and 1019 [-Wformat-truncation=] 118 | "<%i>%.15s %s %s[%ld]: %s", | ^~ ...... 122 | VPREFIX, | ~~~~~~~ ../../libpromises/syslog_client.c:118:17: note: directive argument in the range [-2147483648, 2147483647] 118 | "<%i>%.15s %s %s[%ld]: %s", | ^~~~~~~~~~~~~~~~~~~~~~~~~~ In file included from /usr/include/stdio.h:867, from ./../../libntech/libutils/platform.h:72, from ./../../libpromises/syslog_client.h:28, from ../../libpromises/syslog_client.c:25: /usr/include/bits/stdio2.h:67:10: note: '__builtinsnprintfchk' output 11 or more bytes (assuming 1059) into a destination of size 1024 67 | return builtinsnprintfchk (s, __n, __USE_FORTIFY_LEVEL - 1, | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 68 | __bos (__s), __fmt, __va_arg_pack ()); | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ CCLD logging_test /usr/bin/ld: ../../libntech/libutils/.libs/libutils.a(logging.o):/home/service/.cache/yay/cfengine/src/cfengine-3.15.1/libntech/libutils/logging.c:35: multiple definition of `VPREFIX'; logging_test.o:(.bss+0x0): first defined here collect2: error: ld returned 1 exit status make[4]: [Makefile:1450: logging_test] Error 1 make[3]: [Makefile:2775: check-am] Error 2 make[2]: [Makefile:489: check-recursive] Error 1 make[1]: [Makefile:708: check-recursive] Error 1 make: *** [Makefile:997: check] Error 2 ==> ERROR: A failure occurred in check(). Aborting... error making: %!s(func() string=0x55e3fe465e80)

ektich commented on 2020-05-07 14:53 (UTC)

Thanks TheGoliath, I've updated the URL.

vn158 commented on 2017-05-11 07:52 (UTC)

It looks OK now. Thx.

ektich commented on 2017-05-04 14:25 (UTC)

I've pushed version 3.10.1-2. it at least builds, and passes unit/acceptance tests, but I don't have an environment where I can test that it behaves properly. reports are welcome

ektich commented on 2017-05-04 08:11 (UTC)

I can confirm package is not building. Working on the resolution for this.

vn158 commented on 2017-05-03 10:39 (UTC)

Compiling is not working, error appears during make DONE: Configuration done. Run make/gmake to build CFEngine Community. Making all in libcompat CC generic_at.lo CC closefrom.lo CC rpl_ctime.lo CCLD libcompat.la ar: `u' modifier ignored since `D' is the default (see `U') Making all in libutils CC alloc.lo CC atexit.lo CC hashes.lo hashes.c: In function 'FileChecksum': hashes.c:50:20: error: storage size of 'context' isn't known EVP_MD_CTX context; ^~~~~~~ hashes.c:50:20: warning: unused variable 'context' [-Wunused-variable] make[3]: *** [Makefile:613: hashes.lo] Chyba 1 make[2]: *** [Makefile:487: all] Chyba 2 make[1]: *** [Makefile:677: all-recursive] Chyba 1 make: *** [Makefile:578: all] Chyba 2 ==> CHYBA: Do#lo k chyb# v build(). Ru#ení...

ektich commented on 2017-05-03 09:35 (UTC)

@vn158 what are the symptoms you're seeing? not compiling or not running?

vn158 commented on 2017-05-03 07:22 (UTC)

Package looks broken after openssl upgrade to 1.1 - I tried --with-openssl=/usr/include/openssl-1.0/openssl but it looks like make at least partialy ignores this.

ektich commented on 2016-09-20 06:31 (UTC)

Upstream installs all files in to /var/cfengine, and for consistency with other systems and with upstream's packaging (even though upstream is breaking "rules") I am reluctant to move files into /var/lib/cfengine. regarding libyaml I'll check if it is listed as a pre-requisite upstream or not before changing anything.

bidulock commented on 2016-09-19 13:08 (UTC)

cfengine E: Dependency libyaml detected and not included (libraries ['usr/lib/libyaml-0.so.2'] needed in files ['usr/lib/libpromises.so.3.0.6']) Please add libyaml to depends array. cfengine W: File (var/cfengine/) exists in a non-standard directory. cfengine W: File (var/cfengine/masterfiles/) exists in a non-standard directory. cfengine W: File (var/cfengine/ppkeys/) exists in a non-standard directory. cfengine W: File (var/cfengine/inputs/) exists in a non-standard directory. cfengine W: File (var/cfengine/outputs/) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/) exists in a non-standard directory. cfengine W: File (var/cfengine/modules/) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-monitord) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-promises) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-key) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-serverd) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-agent) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-upgrade) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-runagent) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/cf-execd) exists in a non-standard directory. cfengine W: File (var/cfengine/bin/rpmvercmp) exists in a non-standard directory. These should be in /var/lib/cfengine instead of /var/cfengine

taksuyu commented on 2016-04-25 15:39 (UTC)

Works like a charm now. Good job!

ektich commented on 2016-04-25 13:10 (UTC)

3.7.3-2: I've added cfengine3.service and updated other .service files so that starting/restarting cfengine3.service will start/restart other three. (and I am ignoring the cfengine3.service that upstream provides).

taksuyu commented on 2016-04-25 01:25 (UTC) (edited on 2016-04-25 02:16 (UTC) by taksuyu)

I can understand that, but bootstrapping cfengine isn't the same as enabling them after installation. Part of the functionality is that cf-execd will check if the other services are still up to assure conformity for the system and you get repair failures when cf-exec can't restart them through the cfengine bash script in /etc/init.d/cfengine. EDIT: It's actually the cfengine3 service that it looks for on systems with systemd, but the one they provide in the repo calls the bash script in /etc/init.d/cfengine. It manages successfully killing all the management processes and whatnot. Could probably also make a wiki page. I wouldn't mind helping on that.

ektich commented on 2016-04-21 09:35 (UTC)

Arch's policy is not to enable services after installation, that's why CFEngine's own attempt to use "systemctl restart" after bootstrapping fails. (but I am making an assumption that you did not have relevant services running before bootstrapping). I'll try to reproduce your error later today and see what can be done to make it easier to install/bootstrap CFEngine.

taksuyu commented on 2016-04-21 01:46 (UTC)

I'm having problems bootstrapping a local machine to a policy server (http://hastebin.com/basahetiqe.pl) Any ideas? I figure I can just enable them myself through systemctl, but that doesn't seems like it should be the default.

ektich commented on 2016-02-20 14:46 (UTC)

3.7.2-1: CFEngine 3.7 LTS branch, now using lmdb instead of qdbm.

ektich commented on 2016-02-07 11:10 (UTC)

I've prepared git pull request (https://github.com/zizzfizzix/pkgbuilds/pull/8) that switches to lmdb, but the version there is still 3.6.5. I'll try to prepare a new pull request for building 3.7.2 LTS version during the week.

andreas_baumann commented on 2016-02-07 10:49 (UTC)

Latest versions are 3.7.2 LTS or 3.8.1.

andreas_baumann commented on 2016-02-07 10:46 (UTC)

qdbm no longer exists as package or AUR package?

zizzfizzix commented on 2014-04-07 21:30 (UTC)

@ektich Thanks for your patches. As to your issue if I understand it correctly you're just doing it wrong not using systemctl in the first place in your policies. Here's an example how to do this: https://groups.google.com/forum/#!msg/help-cfengine/BGMrLuWb79k/E31PL5FtgXgJ If you drop any "manual" service activation there won't be any clashes with systemctl, and since it's a specialized tool you can't do any better anyway ;)

ektich commented on 2014-03-12 17:56 (UTC)

I've pushed my changes to some of files to https://github.com/ektich/pkgbuilds. Please review and let me know what you think. One issue (probably site-specific): after successful cf-agent --bootstrap, once client receives policy files from the server cf-execd is started automatically. Within 5 minutes cf-serverd and cf-monitord are started as well. Of course systemctl fails to notice them and starts another set of executables if it is executed with "enable" followed by "start" commands. But without enabling at least cf-execd with systemctl client will stop running CFEngine after a reboot. Maybe a README file documenting this should be included with this package?

ektich commented on 2014-03-11 17:59 (UTC)

Not sure if this is a right place to discuss issues, but I haven't found any better place. I run a Debian-based cfengine server, and I need to add few Arch-based clients to it. After downloading this PKGBUILD, building and installing it I ran into two problems, while running cf-agent -B master-servers.name: 1) "error: needs to be installed in /var/lib/cfengine/bin for pre-validation of full configuration". this is easily fixable by creating a symbolic link from /usr/bin/cf-promises 2) the cf-agent -B re-generates inputs/failsafe.cf file with hard-coded path to the promises on the master server (/var/lib/cfengine/masterfiles) and then fails to copy them from the server. On the server side (installed from debian package provided by cfengine) those files are in /var/cfengine/masterfiles. changing PKGBUILD to compile cfengine with --with-workdir=/var/${pkgname} fixes this issue. third issue appears once cfengine is installed in /var/cfengine: cf-agent -B master-server tries to start cf-execd once policy files are copied, but it expects to find cf-execd binary in /var/cfengine/bin (same as cf-promises). Again, can be fixed by simple ln -s They might not be issues, depends how you look at them. But, if this is the proper place to discuss them, wouldn't it be logical to follow CFengine's "way"? (installing everything in /var/cfengine instead of /var/lib/cfengine, and making sure cf-promises and at least cf-execd exist in /var/cfengine/bin, maybe soft-linked)? I need to read up on PKGBUILD etc but I'll try and prepare a version that does all that for your review.

zizzfizzix commented on 2013-11-14 20:38 (UTC)

I have made quite a lot of changes, so be sure to review them while updating

andreas_baumann commented on 2013-06-20 10:51 (UTC)

You are absolutely right, fixed, thanks.. :-)

zizzfizzix commented on 2013-06-20 10:27 (UTC)

Hey, thanks for updating. Shouldn't it optdepend on libmariadbclient instead of mysql? The latter has been dropped.

andreas_baumann commented on 2013-02-13 19:49 (UTC)

Anyway.. I'll read more, because as I stated earlier, I'm a systemd beginner. :-)

andreas_baumann commented on 2013-02-13 19:40 (UTC)

Well, it's in the testing phase. I would like to use cfengine for an Arch-based infrastructure, so far, I didn't have the time. Concerning -F: I think running in the foreground is for testing and debugging. cf-* are Unix daemons which should run once and only once and hence have a PID file. Ok, the PID file is in a non-FHS-conform location in '/srv/cfengine' but this is something which should be fixed upstream.

zizzfizzix commented on 2013-02-13 17:11 (UTC)

@andreas_baumann Thanks for adding the service files! Can you justify choosing Type=forking instead of specifying -F cli option for executables? I honestly can't see which option is better, but I know -F was adjusted to increase systemd/upstart/etc. compatibility, so maybe it is the way to go?

andreas_baumann commented on 2013-02-13 15:23 (UTC)

Added systemd support. Feedback welcome. I got the daemons up and running (doing nothing) with: cf-key cf-agent -B touch /srv/cfengine/inputs/failsafe.cf touch /srv/cfengine/inputs/promises.cf systemctl start cf-execd systemctl start cf-serverd systemctl start cf-monitord

andreas_baumann commented on 2012-12-05 08:46 (UTC)

I have to dig in and read a lot on systemd, this may take a while :-)

zizzfizzix commented on 2012-11-19 00:04 (UTC)

could you please add systemd service files, e.g. from https://build.opensuse.org/package/files?package=cfengine&project=home%3Adlovasko%3Abranches%3Asystemsmanagement ?

andreas_baumann commented on 2012-10-20 20:03 (UTC)

updated to 3.3.8, new dependency libvirt

andreas_baumann commented on 2012-06-02 07:47 (UTC)

updated to 3.3.3

andreas_baumann commented on 2012-05-16 08:17 (UTC)

updated to 3.3.1

andreas_baumann commented on 2012-04-11 15:40 (UTC)

updated to 3.3.0

fukawi2 commented on 2011-10-29 09:15 (UTC)

Update to 3.2.3 Thanks andreas_baumann, I've updated the note about deps, and the source URL did change slightly ('source_code' became 'source-code')

andreas_baumann commented on 2011-10-29 08:52 (UTC)

source=("http://cfengine.com/source-code/download?file=${pkgname}-${pkgver}.tar.gz") actually works fine. I was wrong here.

andreas_baumann commented on 2011-10-27 03:40 (UTC)

Tried 3.2.3, compiles. There are some issues: The advice to comment out configure options could trouble newbies with messages like: /media/sd/arch/adapted_packages/cfengine/PKGBUILD: line 33: --with-postgresql: command not found Better would IMHO be to just add a comment there and not a message: Arch users should feel comfortable to edit PKGBUILD files anyway. The source can't be downloaded, but that's a cfengine policy for the community edition. Maybe here a message would be appropriate and setting the source value in the PKGBUILD as follows: source=("${pkgname}-${pkgver}.tar.gz") Just an idea. :-)

fukawi2 commented on 2011-09-11 02:28 (UTC)

Urgh, sorry... Blonde moment. Fixed.

andreas_baumann commented on 2011-09-10 08:09 (UTC)

I think 'libmysqlclient' is enough, no need to have the full server as dependency..

fukawi2 commented on 2011-09-09 08:20 (UTC)

Bump to 3.2.0; Thanks andreas_baumann libpromises.so seems to depend on the mysql/postgres libs if you compile with those options, so I've added them as a dep rather than optdep

andreas_baumann commented on 2011-09-09 05:55 (UTC)

Or maybe postgres-libs and mysql, pcre should be optdepends?

andreas_baumann commented on 2011-09-09 05:54 (UTC)

3.2.0 is out. I put an updated version of PKGBUILD to: https://gitorious.org/aba_arch_aur/aba_arch_aur/blobs/master/adapted_packages/cfengine/PKGBUILD --with-sql I would ommit and add the database connectors --with-postgresql, --with-mysql. The cfengine people say --with-sql is deprecated. --with-pcre is maybe also a nice thing. Sadly cfengine has no pluggable module concept for things like database connectors. :-( The source file can't be downloaded and must be put to the directory containing PKGBUILD before starting to build.

fukawi2 commented on 2011-07-21 10:33 (UTC)

Latest version listed as a release on the website; should be no problem to downgrade to 3.1.5 by updating pkgver and md5sums()

andreas_baumann commented on 2011-07-21 09:36 (UTC)

I'm just curious: why do you prefer the beta 3.2.0b1 version over the last stable 3.1.5 version?

andreas_baumann commented on 2011-07-18 11:03 (UTC)

Thanks a lot. :-)

fukawi2 commented on 2011-07-18 10:36 (UTC)

Err, whoops.. Yes, thanks :) Fixed, and bump to 3.2.0b1 https://github.com/fukawi2/aur-packages/commit/f8229f8d8bed569bafc6587062646a8933c1808a

andreas_baumann commented on 2011-07-18 07:18 (UTC)

The make in package should maybe run 'make install' not just 'make', my package is completly empty.. :-)

fukawi2 commented on 2011-06-01 23:10 (UTC)

Thanks, updated to 3.1.5 :)

anrxc commented on 2011-06-01 22:11 (UTC)

Hi I flagged it because it's up to 3.1.5 now.

fukawi2 commented on 2010-09-06 10:42 (UTC)

Bump to 3.0.5p1