Elastic Search Lucene version Issue

Hi All,

I have recently upgraded my Elastic Search cluster from v5.2.2 to v6.1.2.
I did it using cluster restart process as specified in the Elastic Search guide.

I fired the _cat/segments command and noticed that there are 2 Lucene versions visible in the cluster for the same shard. (Screenshot attached)

Can someone help me with this? Is it a common scenario after upgrading?

I have also observed that the performance of the cluster has degraded. Is Force-Merge an option to improve the performance?

Kindly help.
Thanks in advance.

Please don't post images of text as they are hardly readable and not searchable.

Instead paste the text and format it with </> icon. Check the preview window.

Probably you created new data after the upgrade which causes to create new segments.

Is there any problem with that? I mean that having mixed segments should be a big deal.

Hi @dadoonet,

Thanks for the reply.
Yes, after the upgrade, new data has been inserted into the indices and probably, that is the reason for segments with different Lucene versions.

Just wanted to know if it is common to have different Lucene versions in the same cluster.
The cluster is working fine but the performance of the cluster seems to have degraded a little.
Any steps that I must take to improve the performance?

Regards,
Raj Shah

Yes it's common. Calling forcemerge API will probably get rid of the older segments, but it's an intensive IO operation.

Hi @dadoonet,

Thanks.

Is there any other way where we can update the existing Lucene version via some command or process?
And any performance improvement tips for v6.1.2?

Regards,
Raj Shah

Is there any other way where we can update the existing Lucene version via some command or process?

That's what I said: _forcemerge API.

As David said, use the forcemerge API and set max_num_segments to 1.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.