I received a couple HadCM3 tasks a few hours ago and they both failed with repeated segmentation violation crashes within a minute. It appears to be another batch that is failing on Linux and so far working fine on Windows and Mac.

I also received some of these tasks. On one machine which got two, they both failed within 3 minutes. On the other I have four which are now up to 5 hours. I can't remember the details but I think this batch is pushing the envelope in terms of extreme parameters being tested which means a high failure rate is expected. I am sure it isn't anything to do with the machines as the one which has the still running tasks had four of the same type of task fail with segmentation violation last time around. I notice that the tasks are new ones but from old batch numbers so it may be that eventually new tasks from these batches will have a higher success rate as they work out what parameters are worth investigating.