[Triolith-users] File system problem, login node restart tomorrow

[Pär Lindfors]

Dear Triolith users,

This mail contain Triolith specific information about the file system
problem described in my recent mail to all NSC SNIC users. [1]

Compute nodes was upgraded to a version of GPFS affected by this problem
on 2014-09-08. This means that jobs started from 2014-09-08 until
2014-10-01 have been running on the bad version.

Login nodes was upgraded to the affected version during the downtime
2014-09-17.

We will restart login nodes tomorrow (2014-10-03) to apply the fix.

* triolith2.nsc.liu.se will be restarted 2014-10-03 at 09:00.
* triolith1.nsc.liu.se will be restarted 2014-10-03 at 17:00.

Before the restarts all logged in users will be logged out, ThinLinc
sessions and interactive jobs will be killed. Running and queued normal
batch jobs will not be affected.

[1] https://lists.nsc.liu.se/mailman/public/snic-users/2014-October/000042.html

Regards,
Pär Lindfors, NSC