[Snic-users] Slow performance on NSC Centre Storage today

[Mats Kronberg]

Summary: /nobackup/global performance is now back to normal. Jobs that ran
between 01:10 and 10:40 today might have been slowed down due to slow
performance on /nobackup/global.

If you are concerned that your job might hit its wall time limit due to
this, please email the job IDs to support at nsc.liu.se and we will extend the
wall time limit for that job.


Details:

The /nobackup performance problem seems to have started around 01:10 and
things returned to normal around 10:40 while we were still investigating
the problem.

We collected some data, but did not manage to find the exact cause. Our
best guess is that one or more jobs performed disk I/O in some unusual way
that overloaded the storage system or triggered a software bug.

If anyone happened to be testing new exotic disk I/O code between 01:10 and
10:40 today, please let us know so we can see if that was the cause of this.


-- 
Mats Kronberg, NSC Support <support at nsc.liu.se>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.nsc.liu.se/mailman/public/snic-users/attachments/20130702/d91860eb/attachment.html