Parsing Logfile for Errors

[saoirseguita]

Hi there,

A log output from my recent cactus run is attached, along with the corresponding SLURM input file. The error(s?) in the log largely seem to be pointing towards resource allocation and out of memory events, but I'm having some trouble parsing it. I have a few questions:

• Are there other errors apart from memory issues?
• Can I restart this same job?
• Are there any tweaks you could recommend to mitigate these errors? I'm thinking that we could increase the maxDisk / maxMemory values, or just add extra nodes.

Many thanks in advance!

mgaCactus.txt log.txt

[glennhickey]

Please read this before running on slurm -- especially that second paragraph.

From your log, signaled SIGKILL is the important part. This means all your jobs are getting killed by something outside of cactus. It could be because it's using too much memory, but can't say for sure from the log.

You can definitely restart any job using the --restart option.