I have 2 systems dat are attached to Predictor, and both are processing a WU that is stuck on 89.00%. They're at that point for more then 20 hours now (bot machines), and the 'To completion' time is climbing.

I suspended the project as I think I'm wasting time here, the WU will never finish.

Both systems are running BOINC 4.43 and mfoldB125 4.30.

Anybody else with this problem. If so, I'll reset the project (and lose 48 hours of work).

Willy

This is a problem many of us are seeing. There is a thread in the Predictor message board talking about it. This is the reason we are seening new apps almost hourly :roll: The solution is to abort the stuck WUs.

chablis

I've encountered this same processing error today on 3 of my Hosts. Unfortunately, the abort and move on to the next WU is the remedy. Luckily mine were discovered only hours into the crunching. One was stuck at 70%, one at 41% and the other I forgot. Looks like we'll have to keep an eye on PP@H processing status until they fix this bug.

I found this link that explains why the WU's are getting stuck:

[url=http://predictor.scripps.edu/forum_thread.php?id=1729' target='_blank]http://predictor.scripps.edu/forum_thread.php?id=1729[/url]

Looks like you need PP@H to run exclusively or set the run time between switching to a long enough time to allow processing to reach 100%.

From [url=http://predictor.scripps.edu/:' target='_blank]http://predictor.scripps.edu/:[/url]

MFold Problems. There is a bug in our MFold binary that is causing problems on windows machines. If You are running Predictor on Windows and have workunits that are from a version of MFold other than 4.28 please cancel them!