StoreEver Tape Storage
cancel
Showing results for 
Search instead for 
Did you mean: 

Problem with unload of tape

 
Highlighted
Occasional Visitor

Problem with unload of tape

Hi. I have a problem with my new library HP MSL 2024 G3 Series, with a LTO4 drive.

I'm using dp6.10. 

When the backup complete, the library don't unload the tape.

If I see the screen on the library, I see a message SEEK/REW flashing 20 seconds each.

If I execute, from the command line, the command:

uma -ioctl /dev/sg1

I see the tape loaded in the drive.

I can, by hand, with the command move d1 sX, unload the tape.

I don't know if:

a) I have a library problem. I did download the ticket support, but, is a .ltd, and, I can't read it.

b) I have a dataprotector version problem. This version are working with the old library.

c) I have a zonning problem. Is it neccesary zonning using esx4?. I have: the cell manager in a virtual machine, the media agent in a linux virtual machine. I did present the library to the linux machine only.

 

TIA

jorge infante

rosario - santa fe - argentina

2 REPLIES 2
Highlighted
Honored Contributor

Re: Problem with unload of tape

You view the "ltd" support ticket files with the HP Library and TapeTools (LTT) software:

http://www.hp.com/support/tapetools

 

If you could post that ticket here that would give more information to help diagnose your issue.

 

The general response for an issue like this is to get the drive firmware updated to the latest and then use L&TT to erase the tape and following that you may be able to use it again.  This sounds like the drive may be having trouble locating a position on the media during the unload.

Highlighted
Occasional Visitor

Re: Problem with unload of tape

After some research, we're quite sure to have a virtualization issue.

We moved the BMA service (HP DataProtector 6 backup media agent) back to a physical server, and until now 3 jobs were successfully completed.

The FC zoning remains the same, so it's probably not the cause of previous failures. Anyway, our goal is to place the BMA inside a VM (can't afford dedicated hardware for it). In the physical (successful) test setup, the Linux running directly on hardware "sees" 2 libraries and 2 drives -most probably because of the redundant FC path.

The BMA sees the device pairs, and offers a multipath option that we enabled. In the virtualized (troubled) setup, we are presenting the library/robot to all ESX4.0 hosts, one of them with the BMA VM, running Linux. That host "sees" two libraries and two drives but "shows" only one library and one drive to the VM, as SCSI devices. The BMA sees 1 library and 1 drive.

 

Can you suggest a better practice?

Any ideas are welcome.