ElasticSearch replication is not working on the specific couchbase node


#1

Hi all

I have 3 couchbase nodes and configures XCDOR with ElasticSearch server. I have 1.2M docs to replicate. Couchbase had successfully replicated until 400K docs (33%), but stopped replicating with errors after that point.

I checked ‘dcr_errors.1’ log files on all nodes. One of the nodes has been raising this error logs below:

[xdcr:error,2014-07-03T12:36:18.596,ns_1@couch-node3.myserver.com:<0.31860.950>:xdc_vbucket_rep:terminate:390]Replication (CAPI mode) f4268e62702130298bf87f17cc481219/my_index/my_index (my_index/1022 -> http://*****@10.123.123.123:9091/my_index%2f1022%3be99607150f6847cfb927372590a81071) failed: {http_request_failed,“POST”,
http://MyAdmin:*******@10.123.123.123:9091/my_index%2F1022%3Be99607150f6847cfb927372590a81071/_revs_diff”,
{error,{code,503}}}

‘Error code 503’ seems to mean ‘service unavailable’. However, the status of elasticsearch server is very stable. It has low cpu utilisation, low load, and enough memory. I can’t also see any errors from elasticsearch log.

I tried to restart elasticsearch server many times, but it didn’t fix the issue.
I also restarted that couchbase node, but it also didn’t work.
I removed the elasticsearch index and restarted to replicate, but it always stopped after 33%.

My couchbase version is 2.2, and elasticsearch is 1.0.1 with 1.3.0 transport plugin. All of the configuration is default.

Do you guys have any idea to fix this issue?

Thanks,


#2

Hi,
I experienced exactly the same problem and opened an issue on the github project.
I have 2 million documents and that just stopped after 600000 document indexed.
Have you found a solution?