Rollover Automation with ILM

Hi All,

I have face a issue with indices rollover automation with ILM. Created ILM policy with rollover Max index size is 1GB. But rollover is happening at 1.2 GB. almost 0.2 GB lag is there.
Also, changed the Cluster ILM poll interval to 1s.

Any indices performance issue will be face by setting ILM poll interval to "1s" ?
How to reduce 0.2 gb lab on rollover condition ?

There may be some differences like this that arise due to the distributed nature of Elasticsearch.
Though it's only 200MB here, so nothing major, and I wouldn't expect to see that become 2GB if we were talking about a 100GB index.

Size can vary quite a lot depending on how segments have or have not been merged. For a fair comparison always forcemerge down to a single segment.

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