I need some advice with a high load indexing server.
One of our Indexing server is managing indexes for 7 vault stores. The index volume count on this server is over 150000.
So if the index service restarts, it doesnt become fuly functional until the volumes are synchronized. Any everytime the volume sync just stop after a certain point and 41376 events repeats the same nummber without the indexing service becoming fully functional.
One option we have considered is to move half index locations from high load server to a different server from SQL server.
Is this a viable option? If anyone has any other suggestions, please do share!
Of course this is a viable option, but you might also <just> want to diagnose why the service doesn't become fully functional, maybe using DTrace.
So the next time you need to restart the service, try to enable dtrace first, so you capture some logs for further diagnostics.
Also, what do you mean by "a different server from SQL server" ?
I did try to troubleshoot this issue but was informed by support that xml files are limited to 2 GB so repository.xml should not grow beyond that. At the moment, our repository file is almost at 2 GB.
Also, I was told that if the index volume count on a single server is beyond 40K indexing service has issues coping up with that.
I assume these factors have something to do with the sync not completing.
I meant making the change directly in the directory database, IndexRootPathEntry table since our index location is on a CIF share.