Is there a max file count a single forwarder can monitor? I have some oracle applications that generate 10,000's of new files daily, various sizes. I have a dedicated host that has these shared mounts mounted up and just scanning for new data. Every how and again it crashes the box. I can't seem to make anything out on the console.
Am I running into some sort of file limit?
The Splunk on Splunk app will give you an idea of what's causing the queue blockage.
It's possible for a single Universal Forwarder to saturate an indexer if it has sufficient event throughput.
09-04-2012 00:10:36.644 -0500 INFO TailingProcessor - Could not send data to output queue (parsingQueue), retrying...
09-04-2012 00:10:37.836 -0500 INFO TailingProcessor - ...continuing.
For one of the data sources, I have, and it works just fine with about 100,000 files. It's when I added several other data sources it broke down.
I do have these errors in splunkd.log that I've just noticed:
09-04-2012 03:06:06.730 -0500 INFO TailingProcessor - File descriptor cache is full (100), trimming...