Admins

Il y a eu des OutOfMemory sur Opium depuis le 17/07 à 17h. La cause n'en est pas connue.

NRPE était coupé, et la charge de la machine était de 6.7 quand je me suis connecté pour aller voir ce qu'il s'y passait :

 14:59:24 up 16 days, 16:51,  1 user,  load average: 6,71, 6,69, 6,48
root@opium:~# /etc/init.d/nagios-nrpe-server restart
Stopping nagios-nrpe: nagios-nrpe.
Starting nagios-nrpe: nagios-nrpe.

Après redémarrage de NRPE, on a pû voir qu'il y avait eu des OOM d'enregistrés :

root@opium:~# cat /tmp/nagios_oom_killer 
Jul 17 17:00:06 opium kernel: [1364058.055209] Out of memory: kill process apache2(1651:#10) score 199156 or a child
Jul 17 17:00:08 opium kernel: [1364059.984953] Out of memory: kill process apache2(1651:#10) score 196803 or a child
Jul 17 17:00:08 opium kernel: [1364060.003743] Out of memory: kill process apache2(1651:#10) score 136331 or a child
Jul 17 17:00:08 opium kernel: [1364060.018657] Out of memory: kill process apache2(1651:#10) score 58074 or a child
Jul 17 17:00:08 opium kernel: [1364060.038034] Out of memory: kill process apache2(1651:#10) score 58074 or a child
Jul 17 17:00:08 opium kernel: [1364060.134417] Out of memory: kill process apache2(1651:#10) score 33173 or a child
Jul 17 18:00:13 opium kernel: [1367657.841123] Out of memory: kill process apache2(1651:#10) score 103118 or a child
Jul 17 18:00:13 opium kernel: [1367657.858685] Out of memory: kill process apache2(1651:#10) score 100760 or a child
Jul 17 18:00:17 opium kernel: [1367662.193836] Out of memory: kill process BackupPC_dump(22341:#10) score 29612 or a child
Jul 17 18:00:18 opium kernel: [1367662.918329] Out of memory: kill process BackupPC_dump(22341:#10) score 28933 or a child
Jul 18 15:00:10 opium kernel: [1443108.147783] Out of memory: kill process apache2(1651:#10) score 236025 or a child
Jul 18 15:00:10 opium kernel: [1443108.322384] Out of memory: kill process apache2(1651:#10) score 233667 or a child
Jul 18 15:00:16 opium kernel: [1443113.688949] Out of memory: kill process apache2(1651:#10) score 100760 or a child
Jul 18 15:00:16 opium kernel: [1443114.147917] Out of memory: kill process BackupPC_dump(25306:#10) score 50941 or a child