[Vagnekman-users] Vagn weekly status update, 7/12

[Johan Raber]

Dear Vagn users,

We have now completed two phases of data read-out from the damaged
filesystem, data constituting the vast majority of the contents of the
filesystem (~ 27/28 TB). The two phases were 1) data marked as especially
important by you and 2) everything not marked as unimportant by you.
We have now proceded to reading out the rest as well, a process finished by
tomorrow.

After this has finished the rescue plan deviced by IBM will be attempted
and some types of comparison between the data as they look on the rescued
filesystem will be made to what we have read out. According to our plan the
filesystem /nobackup/vagn1 will then be re-created and data will be flushed
back from the read-out files as deemed appropriate. Once we are happy with
this, things should get back to normal and we will start to work on adding
new storage to the system. A well founded time-estimate for reaching normal
state will be presented next monday.

About the state of the data we have read out so far we have only some
pre-incident recorded md5sums to compare against. However, these indicate
that a fairly large percentage seem to be affected but what this in effect
means in terms of qualitative and quantitative damage is difficult to
judge.

As of wednesday last week you have got read-only access (via Samba) to your
home directories on Vagn and this "solution" will at least remain in place
until end of week (pending the above described data-rescue maneuvres) when
we will try to return to the normal read-write setup (via GPFS).

The temporary access we have set up to /home had some unexpected
consequences, likely depending on the unix permissions export, for some
users who were not permitted to login. This should be fixed by now, try
again if you're in this category and send a support mail if it still
doesn't work.

If nothing unexpected happens a new status update will be coming next
monday.

Best regards,
NSC support