Results file does not exist even though results count is non zero

[mlf]

I have a save search that runs at 5 minute intervals and passes it's results to a perl script for additional processing and possible alerting. Every so often, the script 'fails' because the results filename passed to it by Splunk does not exist even though the results count was positive. (it's normally ~80) As the search is monitoring a critical infrastucture process, these lapses are causing issues.

Has anyone else experienced this? Is this a known behaviour in 4.1? 4.2?

I'm currently at 4.1.6 and will be upgrading to 4.2 within the next 45 days.

[fk319]

I have a similiar issue. There are cases where the server gets to busy doing a search it stops indexing. I have had cases where my indexing is gotten behind by 20 minutes and longer.<\p>
We are running 4.1.5 and we are moving from a single server to a much larger cluster of servers, time frame unspecified.