Backup Exec 12.5 throughput slow

[fish]

I am having trouble getting the throughput to get back to its usual speed. The first question will be what changed. The tape library malfunctioned (HP 1x8 G2 Tapestore LTO4) and was replaced, but prior to replacing they wanted many things done. Update BE, Bios updates to tape drive, windows configuratuions changes (thier best practices), and updates run on the server. (Windows 2008, SP 1, 32bit). So, many things changed, but the final result is from the usual 1200MB/min I am getting 600 MB/min. Spent 5 days with HP working on this and they even replaced with another drive, but still not working. (Using their diagnostics, it runs at peak performance)
The final word was talk to Backup Exec, they have done all they can.

So here I am...... Compression is set to Hardware in the job setup. compression is set to on in the tape device properties.

Any suggestions or known issues with HP?

Peace,
Tom

[cmschube]

Hey Tom -

You can try to disable Read/Write Single Block Mode. I tried this and got really good results. And it would make sense since you changed your device. Go to Devices and right click on your tape until and hit Properties. Under Configuration you'll see the two check boxes, uncheck both of them (I unchecked all 4). And see how that works.




See how that works -

Good Luck -

cmschube

[fish]

I will try that right now with a back up

[fish]

It did not help. It did not slow down either..... I thought I would see a performance hit because having it checked marked bypassed the controller and allowed direct communication. 

Thank you for the suggestion. Well, maybe it did slow down a small amount, I just checked it and it is running a 590 average rather than 600 MB/min

Running HP's tests using Storageworks library and test tools, everything runs fine. Our backbone is 10gig, with 1 gig to the servers, HP managed switches, no changes to network infrastructure, or to the target servers.  I guess I will have to call Symantec tech support...... that is right up there with chewing sand on my most favorite things to do :)

I will update if I get a solution.
Peace,
Tom

[Ken_Putnam]

Tom  

cmschube
had the wrong box highlighted in the graphic

As his text says verify that the Single  Block Write option is NOT selected (or the Single Block Read, either for that matter)

[fish]

Thanks, I got it, neither is selected. Still not hitting the speed we need. Daily is taking 27 hours instead of the 13 it usually completes in.
 

[Ken_Putnam]

OK

If you shut down all BackupExec services and do an NTBackup, what kind of throughput do you get?

Don't remember (and can only see one post back in this response window) - have you recently updated drivers or firmware for either the drive or controller?

Local or Network backup?  If network, can you copy a large block of data (xcopy or FileManager drag and drop) with acceptable speeds?

[fish]

I will need to respond tomorrow after my current backup finishes........ I need to get a good set done (I get very nervous when I do not have a good backup).

[Ken_Putnam]

completely understand   |8^)

[fish]

 Truly sorry for the delay. As the season I am in, the tape library got stuck with a tape. Yesterday, I ran a backup using HP Data Protector Express. The throughput was excellent. So, I am now convinced that it is a Backup Exec Issue, not an HP issue. I am open to thoughts and suggestions. I talked to a dear colleague who stopped by recently and recommended using HP drivers to see if the performance will improve. I am going to try that in a moment. 

[fish]

 Scratch part of my last post.
I am going to use the HP Data Protector to backup all my servers, then I will go back to playing. Again I just will feel better with a complete set of backups. We have a decent amount of data to backup, so I may be delayed in responding again. I will update as soon as possible.

[fish · Accepted solution]

 HP replaced the tape library and some of the problems went away....
One of the issues was that it would begin backing up but at a random time it would reset. Turned out that replacing the library corrected the disconnecting of the backup.
The slowness was still a factor on Backup Exec.
I manually updated (remotely updating failed) the clients on the servers. After that I went in and tried the remote client install and it worked on most of my servers (problems on the 64bit server and the primary domain controller). Ran backups and all seems to be working O.K.