Our backup window for workstations is 12:00-2:00 midday. they 
normally complete in about an hour. One of the policies backs up the 
D-drive on Windows workstations using the archive bit.

Today, after most clients had successfully finished, the D-drive 
policy queued all its clients again, which started to backup ALL 
files on the D-drive! I know it was all files because -

1. The clients started transferring data quickly after their job 
became active. Normally several minutes elapse before data actually 
starts transferring.

2. On the clients that finished, the whole disc was transferred.

3. Clients that had successfully completed only a few minutes 
beforehand with the transfer of only a few MB of data, now were 
transferring gigabytes.

The 'Backup Archive and Restore' program shows backups from the last 
full to the present, so I assume the database is OK. Looking at some 
of the drives being backed up, there were not obviously any files 
with their archive bit set. My next job is to recursively list some 
drives to find all files with the archive bit set.

The only change I made is to tick the "allow multiple data streams' 
box for the D-policy while the first jobs were running today.

Does anybody have words of wisdom, or even just conjecture about what 
may have caused this, and more importantly, what I can do to stop 
another full effectively occurring (I cancelled all of the jobs after 
I realised what was happening).

As ever, thanks in advance.
-- 
---

REMEMBER- There is only one certainty about storage systems - they 
*will* fail !
This includes backup and archiving systems.
If you want to be paranoid, make your own backups as well.

For FIL computer support issues, please email

   support at fil.ion.ucl.ac.uk

rather than individual team members

Chris Freemantle
Data Manager
Wellcome Trust Centre for Neuroimaging

Tel. +44 (0)20 7833 7472 reception
Tel. +44 (0)20 7833 7496 direct
Fax  +44 (0)20 7813 1420

Home Page:  http://www.fil.ion.ucl.ac.uk