OSCER users,

REMINDER:

On ***THURSDAY*** June 8, 8am-midnight, OSCER will hold
a scheduled maintenance outage for all our research computing
systems.

---

Before the maintenance, if you submit a job and it
doesn't begin quickly, that might be because jobs won't launch
unless they're guaranteed to end before the start of
maintenance (Thu June 8 8:00am).

In that case, reduce the wall clock time request, like so:

scontrol update JobID=####### TimeLimit=DD-HH:MM:SS

except replacing ####### with the job ID number and
replacing DD-HH:MM:SS with the new, shorter wall clock time
limit.

You might need to do that multiple times to get your job to
launch. (Or the supercomputer might be too busy.)

If your job isn't able to run to completion, then you're
no worse off than if you hadn't tried to run the partial job
in the first place.

But if the job does complete, then you're better off.

Henry

----------

On Thu, 1 Jun 2023, Henry Neeman wrote:

>OSCER users,
>
>On ***THURSDAY*** June 8, 8am-midnight, OSCER will hold
>a scheduled maintenance outage for all our research computing
>systems, including:
>
>* Schooner supercomputer
>
>* OURcloud
>
>* OURdisk
>
>* OURRstore tape archive (current)
>
>* PetaStore tape archive (old, soon to be decommissioned)
>
>We're decommissioning several pieces of network hardware that
>provide top-level connectivity for all OSCER systems.
>
>NOTE: This maintenance outage will occur on a ***THURSDAY***
>(instead of the usual Wednesday).
>
>As always, we apologize for the inconvenience -- our goal
>is to make OSCER resources better!
>
>The OSCER Team ([log in to unmask])