HP 960 Autoloader going offline with Backup Exec 12.5 SP3

[Martin Pawlett]

Hi,

I'm having problems with a HP 960 Autoloader, and Symantec Backup Exec.

Config;

- Windows 2003 R2 SP2 (32 bit)
- Backup Exec 12.5 SP3
- HP Storageworks 1/8 Autoloader 960 connected to the Windows Server via a HP/LSI Logic PCI-x Ultra320 SCSI Card

For the last few weeks, all the tape backups have failed, all B2D backups run fine. The tape backups all go to the HP Autoloader. The tape backups fail at various points before they get to 100Gb.

The tape backups fail with error;

Completed status: Failed
Final error: 0xe00084f0 - The device timed out.
Final error category: Backup Device Errors

For additional information regarding this error refer to link V-79-57344-34032



Application event logs show errors 33152 and 57665
System event logs show error 11

Once the backup job has failed, the autoloader and tape drives disappear from windows device manager, and I have to restart the server and autoloader to get them back.

I'm using the symantec drivers as recommended by Symantec.

I've run the HP Tape Tools, and that comes back fine, as does a 128Gb test backup via the tools.

I'm still trying to work out if its the software causing issues, or the hardware that may be faulty.

I've uninstalled SP3 for Backup Exec, and rolled it back to SP2 as there have been some posts on other forums about this.

Any more ideas on what to try next?

thanks
Martin

[TapeDrive Killer 1 2]

What driver are you using for HP/LSI Logic PCI-x Ultra320 SCSI Card? Also you should have a terminator plugged on the top SCSI port on your autoloader.

Pablo Alvarado Siles

[Martin Pawlett]

Hi,

I'm using the LSI driver, lsi_scsi.sys dated 03/06/2008, version 1.28.3.0

Yes, the terminator is at the top

thanks
Martin

[Martin Pawlett]

Hi,

Can anyone else help?

I've tried a new SCSI card in the server (374654-B21) and still get the same error.

I've run the backup exec tracer application, and it shows these errors when the backup jobs fail.

The Driver Result section also reports "STATUS_IO_DEVICE_ERROR", that's the tape drive throwing an error.
All the events after that are showing "STATUS_DEVICE_NOT_CONNECTED", which means the SCSI BUS does not see either device anymore.

The tracer programs pulls a trace of everything going across the SCSI bus, so this is outside of Backup Exec.

Any ideas?

[TapeDrive Killer 1 2]

Use LTT to extract a suppor ticket and attacht this here.

Pablo Alvarado Siles

[Martin Pawlett]

Hi,

File attached

thanks

[Bryan sherman]

Hi Martin,
I have been having the exact same issue as you for the past several weeks. While I still don't have the answer, I can share some of what I have tried that might help us to narrow the issue down. Please update this thread if you find a solution and I will do likewise.

Config:

- Windows 2003 R2 SP2 (32 bit)
- Backup Exec 12.5 SP3
- HP Storageworks 1/8 Autoloader 960 connected to the Windows Server via a HP/LSI PCI Ultra320 SCSI 2000 series Card

Description of problem:

My issues are exactly as you described yours in the first post, although I can occasionally get a successful backup. Some of them can get to 20, 48, 200GB and fail, others have succesfully backed up as much as 800GB to two tapes (we have a large number of complex images that we backup). When the job fails all logs on the server (Eventlog, BackupExec error reports, tracer.exe logs) point to the autoloader being suddenly disconnected. The autoloader itself never once has reported an issue. In fact it is sitting there with the tape loaded in a "Ready" state, but as you mentioned even Device Manager no longer knows it is there.

What I have tried (that has not yet eliminated the error consistently):
1) Cleaned the heads
2) New media
3) Updated all drivers (SCSI, autoloader, and the Ultrium LTO3 drive) and I have tried old and new HP drivers in addition to the Symantec drivers for 12.0 and 12.5.
4) Updated the firmware.
5) Swapped the SCSI cable.
6) Removed antivirus software from server.
6) The original problem started on a server running version 12.0, so 12.5 SP3 is not the culprit.
7) Installed SCSI and autoloader on a new server with a completely fresh install of 12.5.
8) I have run the HP L&TT utility and found nothing wrong with the connection, drive or the media. That tool was able to write a full tape without an error.

Tip:
By the way, rather than turning the autoloader off and back on and rebooting the server, try the following easier path to get it running again. Using the front panel of the autoloader select go offline-configuration-reset. On the server use the Backup Exec Services in the tools menu to restart all of the Backup Exec services. This will usually get me back up and running.

It is hard to blame it on the autoloader hardware when the L&TT finds nothing wrong with it, but I am running out of things to try.
Hope something in this post will be helpful to you. Good luck and please keep me posted. I will let you know if I come up with anything.

Thanks,
Bryan

[Martin Pawlett]

Hi Bryan,

Like you, I can occasionally get a successful backup where theres only say less than 50Gb to backup. We have 5 different tape jobs which backup various types and amounts of data (Images, SQL, Exchange, Data etc).

I have the exact same config as you, but yesterday I manage to build a new server and decided to put Windows 2003 R2 Enterprise 64Bit on (as I saw something on another forum that the 64bit o/s didn't have the same issue). I installed Backup Exec 12.5 with no updates. This server had a new SCSI card in too. I ran 2 large backups, and each failed at 6Gb! I then ran liveupdate a few times until it was patched up to date, and the backup job still failed with exactly the same error.

So I've now ruled out;
- SCSI Card
- Cable
- Server

The autoloader never says there an issue on the front panel.

So by process of elimination - unfortunately I think its the autoloader thats at fault. Its no longer in warranty though - so it looks like we'll have to find £2500 for a new autoloader!

ps: thats for the tip about retstarting the autoloader

Martin

[Bryan sherman]

Hi Martin,
I don't know if you have given up yet or not, but I wanted to mention that I have my backups working again. The first thing I changed was to restrict the files being backed up to just our key data files which meant avoiding the C drive almost entirely with the exception of a few key files. It is not a size issue though because this still left me with 100s of GB to back up. The other change I made was to set the compression from "hardware if available" to none. The L&TT said that nothing was wrong with the compression, but I turned it off anyway since this problem seems to be related to a timeout of some sort. Anyway, since making the two changes (and I am not going to mess with it to figure out which one made the difference) I have gotten multiple full and differential backups. Hope this helps.

Bryan

[Martin Pawlett]

Hi Bryan,

Thanks for the additional info. I've ordered a new autoloader which should arrive today.

However I did try your suggestion of disabling compression as we were only getting 1.3:1 previously using the option "hardware, otherwise software" in BackupExec. The job still failed with the same error.

Thanks for all your help, and I'll let you know how I get on

thanks
Martin