Index Size & Stats showing incorrect

[mvenkat_in (Venkatesh)]

Hi
I have setup a new ES instance and created few indices and loaded about 1.8 million docs.
However I am not able to see the size and stats of the indices either in marvel or with Index Stats API.
The hard disk space in the server seems to be relative to the docs volume.

I am confused and Pls let me know what could be the issue and how can I "refresh" the stats.

Thanks

[jochen_st]

Is your marvel-agent plugin actually running?

Our installation does not show any difference between Marvel and the Index Stats API.

[mvenkat_in (Venkatesh)]

Yes. It is running. But If you see the output in the "black screen" which is actually a CURL output, that is also showing wrong. Pls refer "e2e-2015-12-02" Index size - It is showing only 192.5kb for 1.8 million documents. The actual size in the HDD is 398 MB.

After I restart the Elastic Search , the stats are coming normal. Not sure what was the issue.

Post Restart the Marvel Index status is showing "Red" and marvel is unable to start in Kibana..

By the way I killed the elasticsearch with "kill -9 PID".

So In summary
1 - The Indices stats were not showing correctly when ES was running fine
2- After I killed and started ES, the marvel Index is showing red with following exceptions. I am suspecting the index might be corrupted as I stopped ES with Kill -9

"[31]: index [.marvel-es-2015.12.03], type [nodes], id [null], message [ProcessClusterEventTimeoutException[failed to process cluster event (create-index [.marvel-es-2015.12.03], cause [auto(bulk api)]) within 1m]]"

How Can I recover !!