The Scottish Boinc Team

Posted: **Sun Jun 29, 2008 1:30 am**

I just spotted this on the Rosetta home page and then found that one of the duff WUs has indeed been wasting a core.

As many of you have likely noticed, our throughput has decreased significantly in the last week. This was due to a couple bad batches that were sent out last week for CASP8 that were not caught by our Ralph tests. The issue was rather insidious in that the tasks would cause some clients to stall and sit idle so we were not immediately aware of the problem. The task names started with "t405". If you have any of these tasks on your client, PLEASE ABORT THEM. These batches have been cancelled. We are currently testing a fix on Ralph and an update will soon be posted. Sorry for any inconvenience.

So, now you know, act.

Posted: **Sun Jun 29, 2008 6:53 am**

Ah, that'd explain why my Linux machine was messing up about a week ago. I noticed that the manager was frozen a couple of times with Rosetta, so I flushed its remaining jobs and suspended the project for several days.

Peace of mind to know that something was wrong on their end and not here - I hadn't checked their homepage to see the notice. Thanks, Mr. Megacruncher.

Posted: **Mon Jun 30, 2008 3:34 pm**

answers a lot about the bugs i've been having. been getting stalled wu's on my work machine and had no idea why.

i was aborting them anyway.

The Scottish Boinc Team

Please Abort Certain Rosetta Tasks

#1 Please Abort Certain Rosetta Tasks

#2

#3