cancel
Showing results for 
Search instead for 
Did you mean: 

Backup Exec 2012 - Tape backup is really slow

Marco_Tafur_-_Q
Level 4

Hi guys!

Well, we have this customer who just upgraded to BE 2012 (updated to latest patches) to a new box.

Setup is as follows:

----------------------

BE:

W2K8 SP1

BE 2012 (patched to SP1, rev. 1798).

HW is a HP dc7900cmt (i was against this..but they won't pay for a new server and I'm thinking this is related).

LSI Logic-based 8 Port PCIe SAS Adapter (1068E) (latest drivers).

Gigabit Ethernet
-------------------
Tape Drive:

IBM TS 2900 autoloader w/LTO3 SAS Drive (updated to latest FW).
-------------------

So, today I started to launch the jobs and it is really slow (like 500MB/min). So, a 600 gig backup took 30 hours, which is too much. I noticed that BE indicates there are 4K of soft errors and I wondered if this is related to the performance.

To let you know, I have asked the customer to update all drivers on server, clean the drive head, if possible change the tapes, defragment the file server and patch W2K8. 

In resume:

- Do soft errors are directly related to backup performance?

- How can I get rid from those?

- Do you have any insights?

This performance / errors didn't happen with the old box and using BE 2010.

thanks for the help!

 

Marco

9 REPLIES 9

CraigV
Moderator
Moderator
Partner    VIP    Accredited

Hi,

 

Soft/hard errors would indicate a problem with either the tape drive or the tape itself. Get hold of HP Library & Tape Tools and run a diagnostic check against it to rule out problems on the drive.

You can also check out a B2D job and see how fast this is.

If this is a new implementation, make sure that no AV on the server is actively scanning the BE services, and if so, put in exclusions for this.

THanks!

pkh
Moderator
Moderator
   VIP    Certified

When you have an error there are quite a few retries, so it will definitely slow down performance.  You got to check whether the the errors are occuring on one tape or on all your tapes.  If it is the latter, then it is probably your tape drive that is the problem, especially if you use new tapes and still get errors.  If it is the former, then it is probably the tape that is no good and should be replaced.

Sush---
Level 6
Employee Accredited Certified

Hello Marco,

    you may also want to take a look at the following technotes:

 

http://www.symantec.com/docs/TECH24414 : How to troubleshoot issues with a Robotic Library (autoloader/changer) and/or Tape Drive(s)

 

http://www.symantec.com/docs/TECH8326 : Reasons why the data throughput rate can be slower than the theoretical maximum when backing up to or restoring from tape media or disk (B2D) & how to troubleshoot or improve backup performance

 


And for you main questions around the Soft errors please take a look at the following technote;

http://www.symantec.com/docs/TECH7363 : An explanation of "soft" and "hard" write errors visible on the Statistics tab within Backup Exec.

 

Thanks,

-Sush...
 

Marco_Tafur_-_Q
Level 4

Hi all!

I really appreciate your quick replys.

@Craig: We have an IBM Drive and I got ahold the maintenance software. I'll do the test today. There's no AV agent on the BE Server. B2D will be a test, too.

@pkh: It is still doing the backup (verify stage). I has started on TUE at 18.00 hours....today's THU, 8am local and still going. I noticed there are now 8K soft write errors on the tape, and about 2K soft read errors on the same tape after verify and is happening on all tapes. This fact strongly suggests the tape drive is coming to it's final days and must be replaced.

@Sush: I have checked those TN, too. Drive is on HCL and is correctly configured (the same config as the old box). We are using Symantec's drivers.

Tracer.exe will be executed today, after all jobs are finished...that will help to support the tape-drive-replacement theory .

Thanks for all your help! 

Marco

 

Marco_Tafur_-_Q
Level 4

Hi guys,

 

So I got rid of the soft errors with a Tape Cleaning and changing tapes.

However, the slowness persists. Seems like is related to the network interface...We have followed everything on the Technotes.

Any other insights?

Thanks,

 

Marco

CraigV
Moderator
Moderator
Partner    VIP    Accredited

...what do the network and CPU/RAM look like when a backup runs?

Giovanni_Di_Ren
Level 3
Partner Accredited

Hi,

I've the "same" problem..

Same autoloader (IBM TS2900) after a cleaning warning, speed decrease from 3000 MB/min to 500 MB/min. I think is an HW problem....

Usage statistics:

bytes written 205TB

hours in use 1871

soft read error 100

hard read error 2

soft write error 44

hard write error 0

What do you think?

Thanks,

Gianni.

Sush---
Level 6
Employee Accredited Certified

Hello Gianni,

    With the Hard and Soft Write errors even I think its Hardware error. Have you checked with your hardware vendor?

 

 

soft read error 100

hard read error 2

soft write error 44

hard write error 0

 

Thanks,

-Sush...

SebastienL
Level 2

The LTO3 spec gives 80 MB/s max, so 4800 MB/min, but you have to feed it data quite fast, or the tape will turn faster than it writes, and will have to go back before continuing.

You can run a backup of the local machine to tape, and see if the network is related to the problem.  If you have a server-grade NIC, it may be better than the integrated gigabit NIC.  Or try upgrading the firmware of the integrated NIC.

You can also play with the compression and see if you get better results.