cancel
Showing results for 
Search instead for 
Did you mean: 

Backups slowing down again

GreedyGreen
Level 3
Further to another discussion I've posted can I ask for any further advice on slow running backups.

I'm running an IBM x255 with a Ultra160 SCSI supporting 2 LTO3 drives (last post I had said Ultra320, but actually that was running the HDs not the Tapes) I've got Block size at 64k
Buffer size at 1024K
Buffer Count 10
High Water Count 0
Read single block mode - unchecked
Write single block mode - unchecked
Read SCSI pass-through mode - unchecked
Write SCSI pass-through mode - unchecked

If I check the Write single block or SCSI pass-through mode it makes no difference although for some reason that's the settings we have setup on every other server which are running OK, but I've seen a few entries on this frum saying they should be unchecked
I've also tried turning of compression and also altering the buffer size to both 256K and 64k, with no effect.

The server is running Win2003sp2 and Backup Exec is version 10d

Last night I reinstalled Backup Exec and ensured it was patched with Service Pack 4, but the backup starts off and manages to peak at about 750MB/min and then justs gets slower and slower as it runs so that now it's showing a rate of 445MB/min 10 hours after starting. The bulk of the backup is about 800GB of Notes mail files which are all fairly chunky files.

I've one Tape drive installed internally and one in an expansion unit, but have tried having the backup start on either drive with no change to speed.

Any further ideas appreciated.

The previous forum entry can be found at the following link, but bear in mind some details have changed since as mentioned above:
https://www-secure.symantec.com/connect/forums/slow-backup-large-files-lto-drive

4 REPLIES 4

GreedyGreen
Level 3
I have a few event log errors that seem to be relevant. The first two errors have been happening on the server for ages for Tape0 and have poped up for Tape1 since added the second LTO. The third error only seems to have started since the second LTO was installed but more than that I think it coincides with an update to the RAID firmware on the Hard Drive SCSI RAID card (ServerRAID 5i) or the Firmware update on the LTO drives, but I couldn't swear to the timing.
All the errors only occur within a few minutes of the server booting. And the backup speed has been faster than it's currenlty running at even when these errors have been occuring so I'm not sure how relevant they are.
imagebrowser image
imagebrowser image
imagebrowser image

Ken_Putnam
Level 6
From the screen shots, it sure looks like a driver or firmware problem

schedule some weekend time and try upgrading one at a time to current (or backing them out) and run a test after each re-boot

GreedyGreen
Level 3
New update, I've scrounged an Ultra320 SCSI card and replaced the 160 one.
When I started the backup it was still slow through the system files, but started climbing in speed as it hot the Notes Mailfiles. Thought I'd got it solved but as soon as I start up Notes, the speed starts to plummet again.
I know it's competing for resourses and disk access etc, but I don't understand why it effects this server so much. Task manager shows utilisation hovering arround the 10% mark with occasional spikes to 25%, but nothing too exceptional.

GreedyGreen
Level 3
Turns out there's a mismatch between the BIOS/Firmware and the driver versions on the ServerRAID card. I think this is causing the disk access to be too dire for both Backup and Domaino to be using the drives at the same time.

My only problem now is that the firmware update fails....but not a problem for this forum.