I am looking to troubleshoot performance problems (Indexing latency) on my splunk indexers. I noticed that there is a lot of blocking, apparently due to the max_size_kb being reached on the indexqueue. I am not seeing any disk latency, so not sure why this is happening..
Anyone know how the max_size_kb is determined, and if it is changeable?
Yes, it's in the queue settings of $SPLUNK/etc/system/local/server.conf (see http://docs.splunk.com/Documentation/Splunk/4.3/Admin/Serverconf for reference).
Yes, it's in the queue settings of $SPLUNK/etc/system/local/server.conf (see http://docs.splunk.com/Documentation/Splunk/4.3/Admin/Serverconf for reference).
This is answered here: http://splunk-base.splunk.com/answers/38218/universal-forwarder-parsingqueue-kb-size