On killing cbScanpyImport due to Out of memory, process exits with 0 exit code

[pcm32]

When running cbScanpyImport on processes where you are constrained by memory (ie cloud VMs), the OS killing the process due to OOM ends up with cbScanpyImport exiting with a zero exit code. This means that workflow environments will think that the process was succesful when it wasn't.

I suspect there must be some handling of SIGTERMs sent from outside where this is catch and the process doesn't produce an error exit code.

[maximilianh]

Interesting. I can understand that this is a problem. I just checked and cannot find any "try" statements around cbImportScanpy, so I *think* I'm not catching these errors, so I'd expect them to lead to a non-zero exit code... Do you happen to have the full stdout and stderr of one of these crashing processes?

…

On Wed, Jul 5, 2023 at 11:16 AM Pablo Moreno ***@***.***> wrote: When running cbScanpyImport on processes where you are constrained by memory (ie cloud VMs), the OS killing the process due to OOM ends up with cbScanpyImport exiting with a zero exit code. This means that workflow environments will think that the process was succesful when it wasn't. I suspect there must be some handling of SIGTERMs sent from outside where this is catch and the process doesn't produce an error exit code. — Reply to this email directly, view it on GitHub <#263>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACL4TNFY2W5ZFNES2BBONLXOUWGHANCNFSM6AAAAAAZ6VEARM> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>