[Gimle-users] DONE: Short downtime for filesystem /nobackup/fouo1 on Monday, 15-16 CET

[Kent Engström]

kent at nsc.liu.se (Kent Engström) writes:
> On Tuesday, when we got /nobackup/fouo1 working again, we had to move
> the disks in the failed server to a spare chassis that is not part of
> the Accumulus set of servers.
>
> Now, when the original server has been repaired, we need to hand back
> the borrowed server. 
>
> After talking to Anders Höglund, we have scheduled this for Monday Feb
> 28, 15-16 CET. If there are no complications, you will only see accesses
> to /nobackup/fouo1 hang for some time while we move the disks back to
> the original server.

The good news is that the swap has been done and the filesystem is up
and running again.

The bad news is that it seems that at least some jobs using
/nobackup/fouo1 during this have received errors, instead of just
hanging on I/O as we hade expected.

We apologize for the inconvenience this have caused,
-- 
Kent Engström, National Supercomputer Centre
kent at nsc.liu.se, +46 13 28 4444