Modifying VisIt default timeout

[yuheng-guo]

Hi,

I am using VisIt 3.1.2 on a supercomputer to create visualizations for large simulation data. I've been constantly getting this error:

0: ENGINE exited due to an execution timeout of 30 minutes. (Alarm received)
VisIt: Error - The compute engine running on c202-012.frontera.tacc.utexas.edu has exited abnormally.

I searched the VisIt documentation and found that there is a execution timeout that was set in default to 30 minutes. So I guess one of my python scripted commands ran for more than 30 minutes and caused the VisIt engine to quit.

Instead to paralellize my jobs better I want to first explore ways to extend this timeout, but I wasn't able to find a solution. So any suggestion of commands I should try or websites are greatly appreciated!

Thanks

[biagas]

Take a look at our Startup options, specifically -exec-timeout.

[JustinPrivitera]

Is it possible the job is hanging at some stage? Does VisIt time out after 30 minutes no matter what or is it only under certain conditions?

[yuheng-guo]

I am running 200+ stills and a third of them will crash due to this issue. We didn't have this issue before, but we recently changed our streamlines xml files to more computational heavy ones (increased max integration steps by a lot) and this issue emerged. Once I switched back to old xml files the issue disappeared. So to answer your first question, yes I believe VisIt is taking too long executing our plot streamlines function. To answer your second question, our simulation data is in hdf5 files. It is always the same iteration numbers that cause VisIt to crash, whereas other iterations are ran without issue.