Hard Error failed to unload tape from drive

[Jason Bischoff]

I'm getting a lot of errors with my LTO 2/20 library in unloading media.

It happens to both drives in the library, and provides and error:
Failed to unload tape from drive , Drive communication error (63, 105)

The only way I seem to be able to get my drive bak, and unload it, is to power cycle the library.

I'm using the current firmware for both the library and drives.

[Eugeny Brychkov]

Jason,
some things to check:
1. SCSI cables and terminators. Try using HP ones. Please read what's written on the terminator and post it here;
2. which HBA do you use? See http://www.hp.com/cposupport/information_storage/support_doc/lpg50063.html ;
3. library backend SCSI busses should be terminated - if SCSI connector is not in use - terminate it;
4. use HP LTT http://h18006.www1.hp.com/products/storageworks/ltt/index.html to generate support ticket and attach this ticket to your next reply zipped
Eugeny

[Curtis Ballard]

Drive communication error indicates a problem with the communication between the drive and the tape library controller. It is possible that it can be caused by a bad SCSI connection if that connection causes the drive to hang completely but if your system is cabled correctly then the drives are on separate busses and you say the problem is happening on both so that makes SCSI problems very unlikely.

What are the exact firmware revisions you are using?

What backup software are you using?

Check to see if there are particular tapes that are involved in multiple failures. The most likely cause of this problem are bad tapes causing severe write errors. When that happens removing the tape solves the problem.

With the any release of firmware you can use the online drive repair feature to take the drive offline then put the drive back online and retry the unload and you should be successful without power-cycling the library.

If you update to the 1.60.0 firmware an "Emergency Unload" option is available. Emergency Unload was designed to be able to unload tapes with physical damage but it should also be able to unload tapes after a drive communication error.

[Marino Meloni_1]

hi
follow all the suggestion you already receive, but check also if this could depend of your Software. In case a reserve mode Scsi command has been issued by Sw, this do not allow any media unload until this condition has been cleared (by SW or Power Cycle).Check in your logs if your BU task are well terminated.
marino

[Jason Bischoff]

Backup software is HP data Protector 5

Firmware is:
C7200 firmware rev is 1600

Drives (LTO) firmware is E36R/OEM2

It happens on different tapes.

It happens on such a regular basis, that all my other backup jobs fail when one session does not unload the tape.

[Jason Bischoff]

I'm also getting these errors:

[Major] From: BMA@ "" Time: 4/25/2003 11:15:24 AM
[90:64] Cannot unload exchanger medium ()

[Michael Tully]

Hi Jason,

I think part of the problem is most likely underlying omniback/DP config.

First make sure that your tape library has the busy drive handling set to 'eject medium' not 'abort' (failed jobs will leave tapes in the drive if set to ABORT)

If your cell manager is 'HPUX' then also you need to do this.

set the kernel parameter 'st_ats_enabled' to 0 (zero)
Have the EMS monitoring for 'dm_stape' turned off.
These prevent SCSI resets being sent from the host. (This also should be done to any HPUX clients as well)

[Jason Bischoff]

I'm in a Windows 2000 environment.

The other thing I notice, is that when I do an interactive job (ie start a backup job manually) it almost never has a problem.

[Curtis Ballard]

That error can only be one of two things: bad tapes or a bad drive. You indicate that it happens on several tapes. It is possible but unlikely that there are several bad tapes.

Reserved drives can not cause that error so don't bother checking that.

You can set OmniBack to not attempt to recover after an error which will reduce the occurance of the error but the underlying cause is a write failure. Until the cause of the write failure is resolved the problem will probably repeat.

Post a zipped LTT support ticket so I can take a look at the drive logs and see if there is a problem with the drive.

[Michael Flegg_1]

I'm getting the same errors using a Dell 128T which is a re-badged 2/20 with Dell firmware and experiencing the same hard errors. I use Veritas BENT v8.6 rev 3878 and the next to latest Adaptec 39160 drivers (I have been suspecting SCSI drivers so far). I also run the latest f/w available for the Dell version of these libraries. I have 16 of them (128T's) and at least 8 have had the problem. I notice in the event logs (all server are W2K SP3) that I usually receive a lot of adpu160m.sys errors around the time of the failure as well as HPC7200 errors and halfinch-VRTS errors.

So far nothing has worked so I will watch this space with interest!