[monolith-users] Monolith status

[Leif Nixon]

Dear Monolith users,

As you know, we have been experiencing storage problems on Monolith.
Three disks in the RAID unit that contains the /disk/global file
system went bad on us in rapid succession. This is an extremely
unlikely thing to happen, and we are working with our vendor to find
out what might have caused it.

Meanwhile, the RAID unit has been replaced, and we are in the progress
of restoring data to it. However, since there is quite a lot of data
to restore and the old RAID unit is running with degraded performance,
this will be a lengthy procedure. The system will be running as usual
during this period, with the restoration procedure running in the
background.

The procedure will look like this:

- Initially all affected users will have empty directories under /disk/global/<username>/.

- The old contents of each directory will be restored to /disk/global/<username>.restore/.

- When a directory has been fully restored, the contents of
  /disk/global/<username>.restore/ will be moved into
  /disk/global/<username>/, and the <username>.restore directory will
  be removed. If there is a filename collision when the restored files
  are moved, the restored file will get a filename suffix,
  ".restored".

So, if your username is x_nscxxx, and there exists a directory
/disk/global/x_nscxxx.restore, your data is not yet fully restored.

As always, if you have any questions, please don't hesitate to contact
us at <support at nsc.liu.se>.

Sincerly,

-- 
Leif Nixon                                    Systems expert
------------------------------------------------------------
National Supercomputer Centre           Linkoping University
------------------------------------------------------------