Created attachment 678382 [details] emerge --info Running test suite with seed 0x6496daef... init/dirTooLong [ OK ] [ 0.00000383 / 0.00000365 CPU ] init/oom heap-fault-delay=1, heap-fault-repeat=1 [ ERROR ] Error: test/integration/test_uv_init.c:156: assertion failed: _rv == 1 (0 == 1) Error: child killed by signal 6 (Aborted) init/dirDoesNotExist [ OK ] [ 0.00002860 / 0.00002820 CPU ] init/dirNotAccessible [ OK ] [ 0.00005597 / 0.00005601 CPU ] init/noSpace dir-fs=tmpfs [ SKIP ] init/metadataOneTooShort [ OK ] [ 0.00042657 / 0.00041869 CPU ]
Created attachment 678385 [details] build.log
Created attachment 678388 [details] test-suite.log
Super, super weird. I can't reproduce this even with a tmpfs partition. But glad to see there's an issue open in upstream about it. https://github.com/canonical/raft/issues/156 Does it work for you with this patch: https://github.com/ganto/copr-lxc4/commit/ea5494791f093c2a65df6b2f26eb601d203830b5 (https://raw.githubusercontent.com/ganto/copr-lxc4/ea5494791f093c2a65df6b2f26eb601d203830b5/raft/raft-0.9.25-Always-skip-init-oom-test.patch) Also how big is your tmpfs partition for PORTAGE_TMPDIR?
Yes, that patch helps. Size 50G. Here's its record from fstab: none /mnt/portagetmp tmpfs size=50G,noatime 0 0
Could you upload your kernel config and I'll try with that. Do you use gentoo-sources, or vanilla-sources?
sys-kernel/gentoo-kernel
Created attachment 678511 [details] kernel config
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=e61d57fc23f8c6248a546e2d228923309d240c46 commit e61d57fc23f8c6248a546e2d228923309d240c46 Author: Joonas Niilola <juippis@gentoo.org> AuthorDate: 2021-01-03 06:56:32 +0000 Commit: Joonas Niilola <juippis@gentoo.org> CommitDate: 2021-01-03 06:57:13 +0000 dev-libs/raft: skip faulty test Closes: https://bugs.gentoo.org/760072 Signed-off-by: Joonas Niilola <juippis@gentoo.org> .../raft-0.9.25-Always-skip-init-oom-test.patch | 31 ++++++++++++++++++++++ dev-libs/raft/raft-0.9.25.ebuild | 4 ++- 2 files changed, 34 insertions(+), 1 deletion(-)
With your kernel config and after booting to 5.4.80 I was able to reproduce this ONCE. It still worked countless of times after. No idea, disabling it doesn't seem too bad. Will keep an eye on the upstream issue.
