going fine on all fronts at the moment

Unfortunately, I can confirm that the issue is not resolved. Here is what is going on:

1) We are seeing the network adapters on the two servers that handle file upload/downloads filling up their cache and then starting to drop packets. These particular adapters have some known issues with the driver under load, but the question we are looking at is why they are backing up.

2) We are doing some problem determination work on our switches that handle communications between servers. Unfortunately the data collection crashed one of the switches and caused a 10 minute outage earlier today.

3) We have connected a network sniffer to monitor congestion and other issues during an 'episode'.

The challenging thing about this issue, is that once the issue starts, the only way to clear it is if we stop all access to the filesystem and let things settle down. Once everything is calm we are able to start everything back up and it performs fantastic for many hours before occurring again. We have not been able to correlate the episodes with any scheduled task (backups, application scripts, etc). In the short term, we are implementing scripts to do this automatically, but that is a workaround until the fundamental issue is identified and fixed.

Maybe Mario can help.

I've been running with a 0.1 day cache on my two core laptop with no problem. I have been running CFSW exclusively for which the WUs are short (under one hour). Might longer running WUs be a cause of the problem with a smaller cache size?

BFH is the tool of choice when computers don't want to cooperate.