Arch Linux User Repository

Binary builds are available:

https://software.opensuse.org//download.html?project=home%3Ajustkidding%3Aarch&package=mongodb

If you have a pre-sandybridge CPU or an ARMv8 sbc, you can use mongodb44

If you'd like to stay in the V5.0 branch, you can use mongodb50

The mongo utility will be removed in future releases. Switching to mongosh is recommended.

mrohnstock, thanks for your bugreport.

I'm increasingly understanding why there is swearing from previous maintainers of mongodb, in the old community git repo.

I've created a new jira issue asking for comment, because the issue you linked to is closed, and I've had them give no response on other closed issues I've had followup questions on.

I'm not sure what to do here. The way they worded the issue you linked to implies to me that it's just a test failure, and not necessarily affect production. That might indicate just disabling the test. But, I'm fairly uncomfortable with releasing effectively untested code. And, I'm increasingly having the impression based on past and current mongodb history that the mongodb source code is incapable of working with external dependencies in the long term, and we may have to abandon the practice of efficiently working with packaged dependencies and just let it compile in everything it needs itself.

After about 18h of compiling the following error appear:

...
[cpp_unit_test:expression_test] 2019-08-22T22:02:49.161+0200 2019-08-22T22:02:49.158+0200 I  -        [main] RegexMatchExpression                     | tests:   27 | fails:    1 | assert calls:          0 | time secs:  0.002
[cpp_unit_test:expression_test] 2019-08-22T22:02:49.162+0200    RegexAcceptsLFOption    DBException: Location51091: Regular expression is invalid: (*VERB) not recognized or malformed in test RegexAcceptsLFOption
[cpp_unit_test:expression_test] 2019-08-22T22:02:49.162+0200 2019-08-22T22:02:49.158+0200 I  -        [main] SizeMatchExpression                      | tests:    6 | fails:    0 | assert calls:          0 | time secs:  0.000
[cpp_unit_test:expression_test] 2019-08-22T22:02:49.162+0200 2019-08-22T22:02:49.158+0200 I  -        [main] TOTALS                                   | tests:  494 | fails:    1 | assert calls:          0 | time secs:  7.515
[cpp_unit_test:expression_test] 2019-08-22T22:02:49.162+0200 2019-08-22T22:02:49.158+0200 I  -        [main] Failing tests:
[cpp_unit_test:expression_test] 2019-08-22T22:02:49.162+0200 2019-08-22T22:02:49.158+0200 I  -        [main]     RegexMatchExpression/RegexAcceptsLFOption Failed
[cpp_unit_test:expression_test] 2019-08-22T22:02:49.162+0200 2019-08-22T22:02:49.158+0200 I  -        [main] FAILURE - 1 tests in 1 suites failed
[executor:cpp_unit_test:job0] 2019-08-22T22:02:49.178+0200 expression_test ran in 8.06 seconds: failed.
[executor:cpp_unit_test:job0] 2019-08-22T22:02:49.178+0200 C++ unit test build/opt/mongo/db/matcher/expression_test failed, so stopping...
[executor:cpp_unit_test:job0] 2019-08-22T22:02:49.178+0200 Received a StopExecution exception: C++ unit test build/opt/mongo/db/matcher/expression_test failed.
[executor] 2019-08-22T22:02:49.180+0200 Waiting for threads to complete
[executor] 2019-08-22T22:02:49.180+0200 Threads are completed!
[executor] 2019-08-22T22:02:49.181+0200 Summary: 105 test(s) ran in 115.98 seconds (104 succeeded, 287 were skipped, 1 failed, 0 errored)
    The following tests failed (with exit code):
        build/opt/mongo/db/matcher/expression_test (17 UNKNOWN)
[resmoke] 2019-08-22T22:02:49.181+0200 ================================================================================
[resmoke] 2019-08-22T22:02:49.181+0200 Summary of unittests suite: 105 test(s) ran in 115.98 seconds (104 succeeded, 287 were skipped, 1 failed, 0 errored)
The following tests failed (with exit code):
    build/opt/mongo/db/matcher/expression_test (17 UNKNOWN)
105 test(s) ran in 116.00 seconds (104 succeeded, 287 were skipped, 1 failed, 0 errored)
    cpp_unit_tests: 105 test(s) ran in 115.98 seconds (104 succeeded, 287 were skipped, 1 failed, 0 errored)
        The following tests failed (with exit code):
            build/opt/mongo/db/matcher/expression_test (17 UNKNOWN)
[resmoke] 2019-08-22T22:02:49.181+0200 Exiting with code: 1
==> Ein Fehler geschah in check().
    Breche ab...

which got already reported as Won't fix in jira. PCRE needs an fix, listed here. I've created a task for the pcre package here.

READ ME - 4.2.0 is here - The good, the bad, and the ugly

Good! 4.2.0 finally supports python3. The dependency on aur/python2-scons is no longer needed.

Bad! 4.2.0 breaks having aur/wiredtiger as a separate package, so it is no longer needed. Upstream may fix this some day, but advised for now they can't be separate. When mongodb is being built, it recognizes it's not using a separate wiredtiger package and compiles wiredtiger's source into part of mongodb, so it's still there, just not separate. This increases the space required to about 260GB. The combined amount of time it takes to build shouldn't change.

Mixed! 4.2.0 breaks most of the tests performed in check() under devtools. Previously, only 8 tests couldn't run within devtools due to systemd-nspawn banning mlock(). Now, it appears hundreds of them require mlock(). It's no longer feasible to discover and disable all of them, and maintain this list between versions, so if devtools is detected, check() does nothing. In my opinion, that's bad news, because I'd like to run the tests. Many users will probably greatly welcome this change, because most of the space and time required to build mongodb is in check().

So, users of devtools will only need 20GB available to build this, and it will complete in about 30% of the time it used to, due to check() not doing anything.

But, direct users of makepkg (including probably most AUR helpers) will now need about 260GB available just to build this.

@ectospasm, you're getting that building mongodb? Do you have python2-scons 3.1.1-1 (latest) installed, or an older version? What are you using to build? I just built python2-scons 3.1.1-1 then mongodb 4.0.12-2 in clean chroots (using devtools) and didn't get the error you posted.

In case this is due to using an older version of python2-scons, I just pushed an update to the PKGBUILD, to explicitly require version 3.1.1 or higher of python2-scons. I purposely didn't bump pkgrel in this push, because there's no need for anyone to re-compile who successfully did before. It's only useful to anyone failing to compile with this error. If you're using any aur helpers, please make sure they update to the newest PKGBUILD with this explicit version dependency.

I'm getting the following, it appears related to python2-scons:

python2-scons ==>; Starting build()...
Traceback (most recent call last):
  File "/usr/bin/scons2", line 204, in <module>;
    SCons.Script.main()
AttributeError: 'module' object has no attribute 'main'
==>; ERROR: A failure occurred in build().
    Aborting...

I've also reported this on the python2-scons AUR page.

Thanks jokke, pushed fix to work with aur/python2-scons renaming of its binaries, done so it no longer conflicts with extra/scons.

Upstream released 4.2.0 a couple of days ago. It causes two compilation failures, which is why I haven't released it. Upstream says they arguably should test building with compilation options to use dependencies as already installed system libraries, rather than obtaining and compiling each of its dependencies on its own. But, as of now, they don't do this before releases.

Upstream released a working patch in my https://jira.mongodb.org/browse/SERVER-42787 which cleanly applies to 4.2.0, fixing an immediate compilation error.

I'm waiting to hear back on my https://jira.mongodb.org/browse/SERVER-42813. Somewhat late in compilation, there's an error along the lines of part of the mongoDB compilation is expecting to have built wiredtiger from source itself rather than using the system library. I am hopeful they will resolve this quickly like with the first ticket, and that this will be the last compilation error. If they don't do so quickly, I'll likely release 4.2.0 which will build at least wiredtiger internally from source, and remove it as a dependency. (At least building all dependencies internally from source expands the space needed to build from 180GB to 260GB, so I'd prefer not to have to do this.)

This would probably already do the trick. Haven't tried yet though

diff --git a/.SRCINFO b/.SRCINFO
index 4226b0e..4efcc77 100644
--- a/.SRCINFO
+++ b/.SRCINFO
@@ -0,0 +1,2 @@
+# Generated by mksrcinfo v8
+# Mon Aug 12 15:45:30 UTC 2019
@@ -4 +6 @@ pkgbase = mongodb
-   pkgrel = 1
+   pkgrel = 2
diff --git a/PKGBUILD b/PKGBUILD
index f88d835..3731d15 100644
--- a/PKGBUILD
+++ b/PKGBUILD
@@ -13 +13 @@ pkgver=4.0.12
-pkgrel=1
+pkgrel=2
@@ -79 +79 @@ build() {
-  scons core "${_scons_args[@]}"
+  scons2 core "${_scons_args[@]}"
@@ -87 +87 @@ check() {
-  scons unittests "${_scons_args[@]}"
+  scons2 unittests "${_scons_args[@]}"
@@ -105 +105 @@ check() {
-  scons dbtest "${_scons_args[@]}"
+  scons2 dbtest "${_scons_args[@]}"
@@ -108 +108 @@ check() {
-  scons integration_tests "${_scons_args[@]}"
+  scons2 integration_tests "${_scons_args[@]}"
@@ -115 +115 @@ package() {
-  scons install --prefix="${pkgdir}/usr" "${_scons_args[@]}"
+  scons2 install --prefix="${pkgdir}/usr" "${_scons_args[@]}"

Still fails to build after python2-scons changed the names of the binaries to include the suffix "2". e.g. scons2. See the PKGBUILD of python2-scons for more info.

EDIT: Outdated as of 4.2.0. Left here for historical purposes only.

READ ME - aur/python2-scons is now required

The new dependency takes a while to compile, sorry.

extra/scons was based on python2 until its new release 4 days ago, 3.1.0-2. Now, it's based on python3. Upstream's build process is incompatible with python3, explaining why the previously succeeding build now fails. The first failure you run into is the SyntaxError: invalid syntax in SConstruct, in comments below.

There's 2 ways forward: 1. Patch in python3 support. 2. Require aur/python2-scons.

I'm going with option #2. Upstream has had merge requests open for several years adding support for python3, and they don't seem to care about it, because no comments have been on it for a year and a half now. See https://github.com/mongodb/mongo/pull/1216 and https://github.com/mongodb/mongo/pull/1195

If it were just a few small changes, I'd consider patching it in. Looking through these pull requests, it's at least hundreds of lines, maybe thousands, they aren't approved by upstream, and they don't apply cleanly to current versions.