Dear HTCondor experts, I just skimmed through all open tickets and did not find a ticket tracking the regression in HTCondor 8.8 that interactive singularity jobs are only running for 3 minutes (i.e. until the "sleep 180" has finished), or actually do not start at all if the /usr/libexec/condor/condor_ssh_to_job_shell_setup is left as it is. Is this issue tracked somewhere and did I just miss the ticket? If not tracked (yet): Interactive jobs in containers are a critical part of our cluster design, so it currently locks us to 8.6.13 on all workernodes and I'd love to see this regression tracked in a ticket. Cheers, Oliver PS: I am fully aware the regression has been caused by a huge improvement of the implementation (which I support!), and I also know the fix is not straightforward (or at least, I am not aware of a straightforward fix).
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature