couchdb docker on cluster-raid dies

[stefan789]

the couchdb docker on the cluster synology system frequently gets killed by the oom-killer after some days, because of a "beam.smp" process:

from /var/log/messages:

Feb 11 17:24:16 nEDM-RAID-II kernel: [605330.172276] Out of memory: Kill process 16624 (docker) score 710 or sacrifice child
Feb 11 17:24:16 nEDM-RAID-II kernel: [605330.180013] Killed process 15009 (beam.smp) total-vm:10286500kB, anon-rss:6005080kB, file-rss:0kB

[mgmarino]

beam.smp is couchdb.
stefan789 [email protected] schrieb am Di., 16. Feb. 2016 um 16:08:

the couchdb docker on the cluster synology system frequently gets killed
by the oom-killer after some days, because of a "beam.smp" process:

from /var/log/messages:

Feb 11 17:24:16 nEDM-RAID-II kernel: [605330.172276] Out of memory: Kill process 16624 (docker) score 710 or sacrifice child
Feb 11 17:24:16 nEDM-RAID-II kernel: [605330.180013] Killed process 15009 (beam.smp) total-vm:10286500kB, anon-rss:6005080kB, file-rss:0kB

—
Reply to this email directly or view it on GitHub
#49.

[stefan789]

For now I tried reducing the number of replication worker processes from 4 to 2. This should reduce memory and cpu load on the system. Don't know if this will avoid the issue or just make it less frequent.

Another option is to restart all dockers every night by a scheduled task.