Indexer restarts every five seconds


I use Version: 4.0.0-4051 Community Edition (build-4051) on CentOS 6 (from RPM). With one empty bucket, I get following error every five minutes (and also /opt/couchbase/var/lib/couchbase/data/@2i seems to be growing with no data in any bucket and no traffic).

Error is:
Service ‘indexer’ exited with status 1. Restarting. Messages:*indexer).listenAdminMsgs(0xc208054280)
/home/couchbase/jenkins/workspace/sherlock-unix/goproj/src/ +0x320
created by*indexer).run
/home/couchbase/jenkins/workspace/sherlock-unix/goproj/src/ +0x5d
[goport] 2015/12/15 01:23:57 /opt/couchbase/bin/indexer terminated: exit status 2

Would have created the issue in JIRA but seems that forum credentials don’t work there.


Could you check and make sure what you have specified under “Settings > Index RAM quota” is set to a value lower than total free RAM available on the node that is experiencing the crash?


Free memory over 3 gigabytes, data RAM quota 1600 MB and index RAM quota 400 MB.


would you mind lowering the index ram quota 256. I wanted to see if we can eliminate the failures with this default setting.


I have the same log error and decrease the index ram quota to 265 mb
but with no effect


Changed to 256 MB, no change, still same error.


Could you share the indexer.log?


It seems to be 24 MB currently, where do you want me to post it? Or does some amount of tail (few thousand rows) be enough?


if you can share the full view, it will help. I am at if you’d like to privately send that over.


Apologies for the late response.
Looks like you have another app taking over port 9102 on the node indexer is running. Could you free up that port for indexer?

2015-12-23T14:14:31.146Z+02:00 [Fatal] indexer:: Error Starting Http Server: listen tcp :9102: bind: address already in use
panic: listen tcp :9102: bind: address already in use



It’s standardized port for bacula. So no.

1 Like