BE 12.5 and D2D2503i Error reading cartridge Data on exporting virtual cartridge

[Georg Schwarz]

Hi everyone!

This is my first post after a long search throught the forums, if I violate any rules my apologies beforehand!

We use an HP D2D2503i as backup target, which emulates an MSL2024 tape library. Connection is iSCSI with a Network of its own (only NIC, Switch (HP1810G-24), Tapelib)
Firmware version is: 101.038
The system is connected to the Server via an NC380T with iSCSI license

We are using BE 12.5 SP3 with all avaliable patches as of Feb 28 2010.

BE runs on a Win Server 2008 x64 platform
We are backing up the server itself, another 2008x64 and a Win2003 file server.

The emulated tape drive is an HP utlrium LTO3-Drive (using symantec driver 5.1.37.0, the and BE-recommended Unknown Media changer driver) Block size and SCSI settings for the tape drive have not been altered and are default values.

The system has already been working for quite some time now.

A few days ago, I tried to copy a virtual cartridge to a physical one and received an error about inconsistent data by the D2D web-interface.

Since I was startled by that I turned on the verify option in BE and surprise surprise: On a random file I get the error:
e00084c8 - "Backup-Gerät ist ausgefallen (eng: somthing like backup device is offline/not available)
In the event log I find the Errors:
33152 - Adamm mover error read failed
57665 - error during a read request (translated)
34113 - Backup job failed (translated)

The error occurs at random during the verify run.

I ensured that:
- all remote agents have the latest versions.
- Symantec drivers are being used
- smaller backups of remote servers run without problems - no problem with verify
- ethernet and IP connectivity is fine!

I also recreated the virtual cartridges

A few pieces of extra information:
The changer is partitioned in 6 partitions, with three slots for the first 5 partitions and one slot for the last (keeping it empty for restores)
Backup takes place according to a partition chosen by the day of week

I read through a heap of documentation but all I could find was information pertaining to physical devices.

Does anyone of the inclined readers have an idea what might be the problem here? Management is slowly getting fussed about the whole issue, because I had to tell them that some other backups might also be toasted as we had verify off.

with many thanks in advance!
best regards
George

[Daniel Bernardo]

you can try to unlock the library from BE and then try to copy .... also yo can stop the services of BE and try to copy ...

regards and good look.

[Georg Schwarz]

Thanks for the answer. In the meanwhile I have opened a support case with HP. It seems that the library itself is experiencing some major issues. The guys are working on it. If anyone is interested, I'll be posting the outcome here....

best, George

[ashraf.khalifah1]

then how hp guys solve the prpblem? I have the same problem

[Georg Schwarz]

Yes, they did! It actually went up to a telcon with the developers of this product. The solution however was trivial. Please DO NOT USE WITHOUT CHECKING WITH HP before.

1) Do a firmware upgrade
2) Delete the RAID-set from the RAID-Controller's BIOS. As the config data and the D2D OS are stored on an USB flash drive, you wont harm the config of your system
3) Reboot tice
4) In the Web Interface klick Rebuild RAID (wait one day)
5) That was it for me!

What they figured out as the initial reason was, that probaly one disk failed, and druing the rebuild another ahd a bad sector!

And yes, do a firmware upgrade, it boosts performance incredibly...

best,
George

[ashraf.khalifah1]

thanks for your reply, but after you did these steps you didnt loss the current data on the d2d, right?

[Georg Schwarz]

Well, sorry I should have been more clear on that... Hope you're sitting ;-)
As the RAID had to be rebuilded from Scratch all the data was lost. We had a dreadful night until the first full backup had completed...