One of the files on the server had it's permissions corrupted, and brought most of the components to a halt. I have resolved the issue, and everything appears to be running fine now.

-Matthew N

With this project in particular, outages like this have an immediacy that other projects don't encounter, the very short queue for GPU's means that within 30 minutes to an hour (perhaps even less) of an outage, there is no work to process.

Other projects which support GPU's offer a much longer queue. I know this issue has been batted around for years (I first started doing work for this project back in the pre-GPU days of 2008), but it still is there.

I guess the end user solution is to always have alternative GPU work available (for me that is Collatz, Moo, and Dnet since I run ATI GPU's).

---