EV11.0.1 Indexing running behind - what to check?

[GertjanA]

Hello all,

We've recently upgraded to EV11.01, which went well.

My EV-servers (3 in total) run fine, except perhaps for indexing. I only do Exchange Mailbox archiving from Exchange 2010.

In the Status Console, I see warnings about indexing. "there are x numbers of items in Vault Store X waiting to be indexed". The number fluctuates a little, but remains around 200.000. Index locations have enough free space. I do not use Index Server Groups. The servers have 16 cores, and 32GB RAM. Site setting has Maximum concurrent index tasks set to 20 (default is 30)

Each index administration task is set to run in the morning from 06:15 to 08:00, and from 18:00 to 23:30

Archiving happens between 20:00 and 24:00, backup is set from 00:30 to 06:00

Daily archiving rate is about 15000/25000 per day per store. I have 4 stores (1 on EV1, 3 on EV2)

I see several events 41380 (Index volume agent failed to start due to time out.) on the EV-servers.

AntiVirus exclusions are set in the AV application used, according to the best practices.

I am not sure what to trace/monitor/try, so I hope someone who is running 11.0.1 for a bit longer can give some advice.

Thanks in advance!

Gertjan

[AndrewB]

how about sql? when everything else seems fine, it tends to be a culprit and for some reason especially after a major upgrade.

[GertjanA]

Hi Andrew,

Thanks. I did not mention that, but SQL is good. We have our maintenance plans setup and configured correctly, they run fine, and according to the DBA's, there are no issues with SQL (indexes, etc are fine)

I'll do some further digging.

Thanks.

[GertjanA · Accepted solution]

Found the culprit.

My predecessor deemed it necessary to set NumWordsToWrite in the registry.

(on KVS\Enterprise Vault\Indexing). After removing this key, indexing keeps up just fine.