Troubleshooting

mongod process "disappeared"
Socket errors in sharded clusters and replica sets
See Also

mongod process "disappeared"

Scenario here is the log ending suddenly with no error or shutdown messages logged.

On Unix, check /var/log/messages:

$ sudo grep mongod /var/log/messages
$ sudo grep score /var/log/messages

Socket errors in sharded clusters and replica sets

If you're experiencing unexplainable socket errors between nodes in a sharded cluster or replica set, you may want to check your TCP keepalive time. The default Linux keepalive time is 7200 seconds (2 hours); we recommed that you change this to 300 seconds (five minutes).

To check your current setting:

cat /proc/sys/net/ipv4/tcp_keepalive_time

You'll usually see this:

To change this setting:

echo 300 > /proc/sys/net/ipv4/tcp_keepalive_time

Note that you must alter the tcp_keepalive_time value on all machines hosting MongoDB processes. This include machines hosting mongos routers, config servers, and mongod servers.

Troubleshooting

mongod process "disappeared"

Socket errors in sharded clusters and replica sets

See Also

Labels

Children (3)