I see see these log messages on a regular basis. There is no distinct pattern. Some nights there might be 1, others 2. Today there were three:
Aug 20 04:23:57 r730-01 kernel: pid 4900 (conftest), jid 968, uid 65534: exited on signal 11 (core dumped) Aug 20 04:28:21 r730-01 kernel: pid 19041 (conftest), jid 915, uid 65534: exited on signal 11 (core dumped) Aug 20 04:30:56 r730-01 kernel: pid 57252 (conftest), jid 915, uid 65534: exited on signal 11 (core dumped)
r730-01 is my main server in the basement. It runs a jail named pkg01, which does my nightly poudriere builds. Those jobs are launched from this crontab:
That UID relates to nobody and matches up with what poudriere does (runs as nobody).
[11:46 pkg01 dan ~] % grep 65534 /etc/passwd nobody:*:65534:65534:Unprivileged user:/nonexistent:/usr/sbin/nologin
[11:28 pkg01 dan /root/bin] % sudo crontab -l -u root PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/games:/usr/local/sbin:/usr/local/bin:/usr/home/dan/bin # use /bin/sh to run commands, overriding the default set by cron SHELL=/bin/sh # mail any output to `dan', no matter whose crontab this is MAILTOfirstname.lastname@example.org 18 4 * * * /usr/bin/lockf -t 0 /tmp/.poudriere.build /root/bin/poudriere-builds.sh
As I type this, I wonder to myself: why not put that in /usr/local/etc/cron.d/? Perhaps another day.
The start of this cronjob aligns with the error messages.
The script being run is:
[11:36 pkg01 dan /root/bin] % cat /root/bin/poudriere-builds.sh #!/bin/sh /root/bin/poudriere-refresh-trees 2>&1 >> /usr/local/poudriere/data/cronjob-logs/poudriere-refresh-trees.log /root/bin/poudriere-builds-132-amd64.sh 2>&1 >> /usr/local/poudriere/data/cronjob-logs/poudriere-builds-132-amd64.log #/root/bin/poudriere-builds-132-i386.sh 2>&1 >> /usr/local/poudriere/data/cronjob-logs/poudriere-builds-132-i386.log [11:36 pkg01 dan /root/bin] %
I recently powered off the only i386 host, so that build is no longer required.
Let’s look at the logs.
Perhaps the timestamps on the log files will help.
[11:40 pkg01 dan /usr/local/poudriere/data/cronjob-logs] % ls -l *.log -rw-r--r-- 1 root wheel 31102 2023.08.20 04:48 poudriere-builds-132-amd64.log -rw-r--r-- 1 root wheel 120 2023.08.20 04:18 poudriere-refresh-trees.log
The refresh script definitely isn’t the cause. It finished almost immediately after it was run.
The first log entry is at 04:23:57 – let’s look into the log file for something at about 5:57 elapsed time (04:23:57 – 04:18:00).
We have this:
... [00:04:44]  [00:00:22] Finished graphics/gtk-update-icon-cache | gtk-update-icon-cache-3.24.31: Success [00:05:57]  [00:01:35] Finished devel/libgsf | libgsf-1.14.50: Success [00:05:58]  [00:02:10] Finished security/gnupg | gnupg-2.4.3_4: Success [00:05:58]  [00:00:00] Building devel/subversion | subversion-1.14.2_3 [00:06:06]  [00:01:30] Finished net/avahi-autoipd | avahi-autoipd-0.8_1: Success ....
Let’s try building those ports and see if that triggers the conftest core dumps.
[11:45 pkg01 dan ~] % sudo poudriere bulk -j 132amd64 -p default -z primary -C graphics/gtk-update-icon-cache devel/libgsf security/gnupg devel/subversion net/avahi-autoipd
Nothing in the logs.
Let’s try some earlier entries and see if we can reproduce the problem.
[11:51 pkg01 dan ~] % sudo poudriere bulk -j 132amd64 -p default -z primary -C print/cups net/avahi-autoipd sysutils/nut net/samba416 net/netatalk3
There it is!
Aug 20 11:56:32 r730-01 kernel: pid 55663 (conftest), jid 1015, uid 65534: exited on signal 11 (core dumped)
I tracked it down to net/netatalk3.
Now, what’s the other port which causes this?
Identifying which port triggered which entry
Given net/netatalk3 started at 00:04:36 into the build (roughly 04:18:36), the only candidate is the message at 04:30:56 (the last one).
[00:04:36]  [00:00:00] Building net/netatalk3 | netatalk3-3.1.15,1
At this point, I gave it. It was 8:30 on a Sunday morning and I lost interest.
The port builds, nothing failed in the build. I’m giving up and will add the message to a list of ignored entries.
Thanks for coming to my TED talk.