Couchdb went down in Dev... AGAIN

[ayoyoness]

The couchdb instance on dev went down... the whole leap plattform throws 503... status says it exited but start doesn't work... until you actually stop it first then start again... NEED to investigate the source of the bug and possibly come up with a fix. See picture below for more

[varac]

and absolutely no signs of a cause in the logs (neither in syslog):

[Thu, 03 Mar 2016 05:31:50 GMT] [info] [<0.18409.47>] 127.0.0.1 - - GET / 200
[Thu, 03 Mar 2016 09:33:37 GMT] [info] [<0.32.0>] Apache CouchDB has started on http://127.0.0.1:5984/

[cz8s]

and it never happens on wazokazi, right?

[varac]

@cz8s yes, i saw this only on dev.pix

aaaand again today on dev.pix, last log entry:

[Tue, 08 Mar 2016 09:54:35 GMT] [info] [<0.522.0>] 127.0.0.1 - - GET / 200

[varac]

first of all, i configured dev1 to use a native systemd unit file for couchdb:

• stopped couchdb
• deployed a systemd unit file to /lib/systemd/system/couchdb.service:

[Unit]
Description=CouchDB Server
After=network.target

[Service]
User=couchdb
Type=forking
EnvironmentFile=/etc/default/couchdb
PIDFile=/run/couchdb/couchdb.pid
ExecStart=/usr/bin/couchdb -r 0 -e /dev/null -o /dev/null -b

[Install]
WantedBy=multi-user.target

• executed systemctl enable couchdb
• restarted couchdb systemctl start couchdb

lets wait and see if and when couchdb goes now the next time

[varac]

i asked micah to include it in the couchdb debian package.