Postby adhodgson » Sat Jan 19, 2008 8:11 pm
Hi,
Further to these posts - I found out that BackupExec was the faulting application which was stopping the snapshot from being released - it seemed that if the backup was still occuring whilst the snapshot was still open, and the script was closing it, it would go wrong, even though BackupExec had no business reading from that snapshot mount anyway!
The upshot of this now is that we have consistant backups of Scalix which are running and taking just over 7 hours to complete. I can cope with this.
We initially were going to put this on a Plasmon FC RAID controller, but we had serious issues with this model. Furthermore, there was no real tools that they provided for doing snapshots or live replication between the arrays, like I have seen on other systems, which is why I went down the local storage route.
We are using SATA with just under 200 users now on RAID-10 in a Dell PE2950 server, and we don't really have any IO issues.
Andrew.