Mhmm.. laut syslog hat tatsächlich der OOM Killer zugeschlagen.
Aber warum?
Code: Select all
Jan 3 10:09:18 km31411-05 kernel: [4754669.010648] oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),cpuset=/,mems_allowed=0,global_oom,task_memcg=/system.slice/mariadb.service,task=mariadbd,pid=24161,uid=109
Jan 3 10:09:18 km31411-05 kernel: [4754669.012351] Out of memory: Killed process 24161 (mariadbd) total-vm:3154768kB, anon-rss:1106792kB, file-rss:0kB, shmem-rss:0kB, UID:109 pgtables:2412kB oom_score_adj:0
Jan 3 10:09:18 km31411-05 systemd[1]: mariadb.service: A process of this unit has been killed by the OOM killer.
Jan 3 10:09:18 km31411-05 systemd[1]: mariadb.service: Main process exited, code=killed, status=9/KILL
Jan 3 10:09:18 km31411-05 systemd[1]: mariadb.service: Failed with result 'oom-kill'.
Jan 3 10:09:18 km31411-05 systemd[1]: mariadb.service: Consumed 41min 26.784s CPU time.
Das einzig auffällige unmittelbar davor war im Zusammenhang mit Clam
Code: Select all
Jan 3 10:09:04 km31411-05 amavis[980076]: (980076-20) (!)connect to /var/run/clamav/clamd.ctl failed, attempt #1: Can't connect to a UNIX socket /var/run/clamav/clamd.ctl: Connection refused
Jan 3 10:09:04 km31411-05 amavis[988806]: (988806-20) (!)connect to /var/run/clamav/clamd.ctl failed, attempt #1: Can't connect to a UNIX socket /var/run/clamav/clamd.ctl: Connection refused
Jan 3 10:09:05 km31411-05 amavis[980076]: (980076-20) (!)connect to /var/run/clamav/clamd.ctl failed, attempt #1: Can't connect to a UNIX socket /var/run/clamav/clamd.ctl: Connection refused
Jan 3 10:09:05 km31411-05 amavis[980076]: (980076-20) (!)ClamAV-clamd: All attempts (1) failed connecting to /var/run/clamav/clamd.ctl, retrying (2)
Jan 3 10:09:05 km31411-05 amavis[988806]: (988806-20) (!)connect to /var/run/clamav/clamd.ctl failed, attempt #1: Can't connect to a UNIX socket /var/run/clamav/clamd.ctl: Connection refused
Jan 3 10:09:05 km31411-05 amavis[988806]: (988806-20) (!)ClamAV-clamd: All attempts (1) failed connecting to /var/run/clamav/clamd.ctl, retrying (2)
Jan 3 10:09:11 km31411-05 amavis[980076]: (980076-20) (!)connect to /var/run/clamav/clamd.ctl failed, attempt #1: Can't connect to a UNIX socket /var/run/clamav/clamd.ctl: Connection refused
Jan 3 10:09:11 km31411-05 amavis[980076]: (980076-20) (!)ClamAV-clamd av-scanner FAILED: run_av error: Too many retries to talk to /var/run/clamav/clamd.ctl (All attempts (1) failed connecting to /var/run/clamav/clamd.ctl) at (eval 99) line 659.\n
Jan 3 10:09:11 km31411-05 amavis[980076]: (980076-20) (!)WARN: all primary virus scanners failed, considering backups
Jan 3 10:09:11 km31411-05 amavis[988806]: (988806-20) (!)connect to /var/run/clamav/clamd.ctl failed, attempt #1: Can't connect to a UNIX socket /var/run/clamav/clamd.ctl: Connection refused
Jan 3 10:09:11 km31411-05 amavis[988806]: (988806-20) (!)ClamAV-clamd av-scanner FAILED: run_av error: Too many retries to talk to /var/run/clamav/clamd.ctl (All attempts (1) failed connecting to /var/run/clamav/clamd.ctl) at (eval 99) line 659.\n
Jan 3 10:09:11 km31411-05 amavis[988806]: (988806-20) (!)WARN: all primary virus scanners failed, considering backups
Jan 3 10:09:18 km31411-05 kernel: [4754668.905772] clamscan invoked oom-killer: gfp_mask=0x100dca(GFP_HIGHUSER_MOVABLE|__GFP_ZERO), order=0, oom_score_adj=0
Und laut Dump haben zwei clamscan Prozesse zusammen mehr Speicher belegt
Code: Select all
Jan 3 10:09:18 km31411-05 kernel: [4754669.006809] [1749652] 119 1749652 232240 209996 1789952 0 0 clamscan
Jan 3 10:09:18 km31411-05 kernel: [4754669.007778] [1749653] 119 1749653 230640 206147 1761280 0 0 clamscan
als der eine mariadbd
Code: Select all
Jan 3 10:09:18 km31411-05 kernel: [4754668.973817] [ 24161] 109 24161 788692 276698 2469888 0 0 mariadbd
Ich dachte ja schon, dass die Kiste hardwareseitig von KeyWeb so designt wurde, dass die Konfiguration von KeyHelp dazu passt und umgekehrt.
Und so richtig herumkonfiguriert hab ich an dem System noch gar nicht. Also: Keine Kritik, nur Suche nach Ursachen und Verbesserung der Situation.
(Ich lese, ich lerne, ich versuche zu verstehen.)
Meine Vermutung und Frage: der OOM Reaper Daemon killt den größten RAM-Nutzer, auch wenn dieser nichts dafür kann?
Und wenn tatsächlich Clam den Ärger verursacht, er war ja schon vorinstalliert und -konfiguriert... was mache ich denn nun?
P.S.
Danke, Markus @mhagge für den Wink zu monit. Ich werde mir das anschauen. Klingt passend.