[Bi-users] [DONE] Downtime for some filesystems on July 5

[Peter Bortas]

The upgrade is now finished a bit early. You are welcome to log in again.

A part of the upgrade was had to be skipped, so we'll get back about a
service window for completing this. On the positive side: Since the
part that had to be skipped was also the part that required the
filesystems to be unmounted on the compute nodes, I didn't have to
kill any jobs.

Regards,
-- 
Peter Bortas, NSC


On 27 June 2016 at 17:44, Peter Bortas <zino at nsc.liu.se> wrote:
> Dear Bi Users,
>
> After having completed the upgrade of all the old /nobackup/
> filesystems it is now time to upgrade the latest generations to get
> rid of some really irritating bugs.
>
> We have decided to do this upgrade Tuesday next week, the 5th of July,
> between 13:00 and 18:00. At the start of that window the login nodes
> will be rebooted to get new filesystem clients, so everyone will be
> logged out. We aim to have the login nodes available again before 14:30.
>
> The filesystems affected by the upgrade are fouo4, fouo5, rossby18,
> rossby19, smhid11 and smhid12.
>
> During this window the filesystems being upgraded will not be
> available. Any processes that have open files on them when we try
> to unmount the filesystems will be killed.
>
> Regards,
> --
> Peter Bortas, NSC