Bricks-level backups slowing down 1

[Fatboy0341]

Single Exchange2k3 server with 600 mailboxes. No mailboxes are any larger than 250MB. Server is fully patched all around.

My full backups currently scream along at around 1 GB per minute...so I am almost positive that I have no problems with the SCSI card or the tape drive.

Lately (the last few days) my bricks-level back ups have been taking 13-14 hours or so - as opposed to only 5 hours or so. There have been no major size increases...we're only talking about 28GB or so. Max throughput tops out at just over 36MB per minute. I usually get 100 MB per minute or better. The job eventually finishes successfully - but I don't like it all of a sudden taking twice as long to finish.

We are running Brightstore Arcserve 11.1 (Build 3110). I've always been a Veritas guy - this is the first time I've ever worked with Brighstore.

Any suggestions?

 

[lebisol]

see where it takes the most time...perhaps it is hung up on courropted mailbox.
bricks are slow by nature and really no need for them with 2k3 recovery and shadowcopy
All the best!

:--------------------------------------:

All around in my home town,
They tryin' to track me down...

 

[strongm]

Brick level backups are the spawn of the devil and should be avoided wherever possible ...

(In my opinion)

 

[Fatboy0341]

Much appreciated for the replies - and we're probably going to get rid of the bricks backups soon anyway.

I know bricks are slow by nature - but to go from 6 hours to 14 hours in less than a week with no visible difference in volume of data or number of mailboxes - I'm just wondering why it might have happened.

Thanks for the advice.

 

[lebisol]

well generally spealking it could be:
-network changes
-back server performance
-courropted mailbox or store
anything in exc logs? anything in backup server logs?
All the best!

:--------------------------------------:

All around in my home town,
They tryin' to track me down...

 

[Fatboy0341]

No errors or even warnings in any of the event, system or application logs.

No corrupted mailboxes or store that I am aware of as of yet.

I increased the number of blocks from 1 to 4 in the Exchange Agent settings (in Arcserve). I'll see what happens tonight when it runs. That still wouldn;t explain why it ran great up until a few days ago - even CA says for the number and size of the mailboxes that 35MB per minute is way too slow - especially when we'd been getting up to 110MB per minute (decent for bricks).

Lastly - what different ways are you all using 2k3 recovery services and shadow copy to eliminate the need for bricks level backups?

JB

 

[lebisol]

> recovery service:

http://www.computerperformance.co.uk/exchange2003/exchange2003_mailbox_recovery.htm

> shadow copy related:

http://www.jsifaq.com/SF/Tips/Tip.aspx?id=9698

http://support.microsoft.com/?kbid=822896

http://www.petri.co.il/how_to_use_the_shadow_copy_client.htm

http://support.microsoft.com/?kbid=841696

All the best!

:--------------------------------------:

All around in my home town,
They tryin' to track me down...

 

[Fatboy0341]

Much appreciated for the information lebisol...\m/



"He who laughs last probably made a backup. He who laughs loudest probably hasn't checked his backups in a while."

 

[lebisol]

no problem,
here is one more option to consider
ftp://ftp.compaq.com/pub/products/storageworks/whitepapers/5983-2303EN.pdf

for most small shops it is an overkill...but could apply to you.
All the best!

:--------------------------------------:

All around in my home town,
They tryin' to track me down...

 

[Fatboy0341]

Thanks chief...



"He who laughs last probably made a backup. He who laughs loudest probably hasn't checked his backups in a while."

 

[Fatboy0341]

Problem solved...

We installed SP2 for Tred Micro ScanMail a few weeks ago, and there is a "Hotfix" for this very problem that is caused by SP2. What aggravates me is that I was on the phone with Trend for about 4 hours during all the troubleshooting and specifically asked their support peopel if there were any nown issues with that service pack. One of their supervisors happened to be looking at the cal records before signing off on them as they were closed and he contacted me and emailed me a link to the hotfix last week. Applied it before the backups started that night and bingo...bricks-level backup back to normal.

Nothing like accurate customer service huh?

Hope this helps if someone has trouble with the same issue.


JB



"He who laughs last probably made a backup. He who laughs loudest probably hasn't checked his backups in a while."

 

[lebisol]

nice work boy,
it is amazing how applications that there to 'protect your system' tend to f it up. Al least they didn't give u symantech's approach to resolutions..."disable your file protection"...hmm lets see...it took me 6 months of begging to get the funds to purchase Nav and my solution is to 'disable it', great!
way to stay on top of it...
All the best!

:--------------------------------------:

All around in my home town,
They tryin' to track me down...

 

[Fatboy0341]

Much appreciated lebisol. I've been down that road with Symantec before...and I'd say Trend is right up there with them as far as their approach to fixing a problem.

Just glad it's fixed and hopefully this will make anyone else who experiences this issue's life a lot easier as far as the solution goes.

JB

PS - Bob Marley rules...



"He who laughs last probably made a backup. He who laughs loudest probably hasn't checked his backups in a while."

 

[lipa]

Hi Fatboy0341
Just seen you post on "Bricks-level backups slowing down"

Can you help me with the name of the "hotfix from Trend" and/or a link to this hotfix.

We have exact some problem here using Trend ScanMail and Veritas 10d.

Thanks
Ptn