Posted on October 10th, 2015 at 10:00 AM
Our CouchDB process kept stopping reporting the following in couchdb.stderr:
heart_beat_kill_pid = 29233 heart_beat_timeout = 11 heart: Fri Jul 24 00:04:19 2015: heart-beat time-out. heart: Fri Jul 24 00:04:29 2015: Executed "/usr/local/bin/couchdb -k". Terminating.
The database was hosted on a heavily loaded VM. HEART_BEAT_TIMEOUT in the /usr/local/bin/couchdb script was increased to 30 secs and RESPAWN_TIMEOUT set to 5 and its been stable ever since.