Tape Engine SCSI Problems -- Have Debug Info 1

[jarhoops]

Hi, I have a StorageTek L180 library with 3 IBM LTO Ultrium-TD1 drives. I am using ARCserve v.9, build 2200. I am getting various E3712, E3719, E6096 errors on two of the three tape drives. Here are the errors that appear in the Activity Log:

E6096 Media Error [ABSL:3050 CMD:Ah TapeName:THURSDAY RID:31049 BARCODE:000001]

E3719 Unable to right to media. (MEDIA=THURSDAY [S/N:000001], EC=UNRECOVERABLE DATA ERROR)

Corresponding errors in the tape log:

[5672] ABSL:3050 CMD:a Returning status[2]
[5672] 01/12 09:52:34 ABSL:3050 [REQUEST SENSE ] 03 00 00 00 40 00 00 00 00 00
[5672] <WRITE >, Sense Data as Follows:
[5672] SENSE ABSL:3050 70 00 03 00 00 00 00 1c 00 00 00 00 0c
[5672] EX SENSE ABSL:3050 00 06 00 40 02 00 00 00 02 00 00 00 00 00 00 00
[5672] SENSE ABSL:3050 Media Error [03]
[5672] EX SENSE ABSL:3050 Write Error [0c, 00]
[5672] ABSL:3050 (SCSI WRITE)
[5672] (03) *** Medium Error ***
[5672] E6012 Write error - head sync error during write (0C 00)
[5672] E6092 Fatal - Uncorrectable error - E6092 - Uncorrectable error


If there are any SCSI experts out there please let me know what you think, or if I need to provide more details. Thanks in advance.

 

[vschumpy]

http://supportconnect.ca.com/sc/redir.jsp?reqPage=search&searchID=TEC342813

 

[jarhoops]

Thanks for the response, but I've been through that document 5 or 6 times... I was hoping someone could shed some additional light on the tape log messages since they are more detailed. Seems to indicate hardware issues... Thanks.

 

[vschumpy]

If you've been through the document then "Media Error [03]" should be a pretty big clue.

Perhaps you could give us the results of some of the tests that were described in the document, which of the tests you did, and what happened when you tried the different things it suggests.

 

[jarhoops]

I cannot find "Media Error [03]" anywhere in this document. Perhaps the link you sent me is for a different document than you're thinking of. In any event, here's some of the testing that has been done:

Verified no conflicting SCSI IDs.
Disabled 2 suspect tape drives and was able to perform backups to other drive without a problem.
Confirmed that SCSI adapter drivers are most current.
Confirmed that tape library and drives have most current compatible and supported firmware.
Enabled RSM and disabled OS drivers for tape drives and changers.
Confirmed all hardware is listed in the CDL.
Checked cabling and termination.
Confirmed most current service pack levels (ARCserve and OS).
Checked system event log (nothing appears).
Ran Device Configuration.
Checked FSC logs on tape library.

Again, the errors appear on two of the three drives. Following recommendations I've found on this forum I have turned the tape log debugging on and have captured the information above. Apparently some SCSI experts could assist by looking at that info.

The steps I will follow going forward are:
Disable HP/Compaq Management Agents
Delete the Tape Engine registry key
Run cleaning 10 consecutive times on each drive
Replace two drives generating errors

If there is anyone out there that can add insight from the logs I've posted please let me know. Thanks.

 

[vschumpy]

I don't think you got the meaning of my post.

The "Media Error [03]" is actually from your own debug log.

Media Error == SCSI Sense 03.

The tape debug logs really are fairly simple, and rarely do you need to analyse the whole sense or ex sense, as most of the useful information actually has what it means in text next to it.

If you disabled the two suspect drives and tried to use the SAME tape that reported the error in the disabled drive in the 'working' drive and it still worked then yes it looks like a hardware defect. You can use the STK SCSI toolbox utility to see whether any hard or soft errors are logged to any of the drives. Obviously if the same tape causes the error on the working drive, then it is fairly safe to say it's a media error as is indicated in the log.

BTW Cleaning drives 10 times is excessive and NOT recommended, it will just wear out the head, once or twice is more than enough, it's not like it's a DC6250 or Travan drive.

Deleting the tape engine key and then re-running device config afterwards is probably your next logical step, there are sometimes some wierd things thrown up when this gets messed up.

 

[jarhoops]

Great! That's some useful information. How do I go about obtaining the STK SCSI toolbox? I saw a SCSI toolbox online for about $5000. Is that what you're referring to. Thanks again for the help...

 

[vschumpy]

If you have the option to remotely manage your library using a web interface then you can get access to it thru this, failing this you may be able to look at the drive logs thru the library front panel. Not sure how useful the info from the front panel is though.

 

[davidmichel]

Here is how it works.
The Tape Engine issued a <WRITE > command and that failed.
So the Tape Engine issued a [REQUEST SENSE ] command to the tape drive to find out what happend.
The tape drive responded by sending its Sense Codes. (The tape drive has registers that maintain data. Part of that data is the Sense Code. The actual Sense Code is byte 2 (counting from 0). In this case it is 70 00 03 which means
Media Error [03]
also included are the Additional Sense Code and Additional Sense Code Qualifier bytes 12 & 13.
Write Error [0c, 00]
The Sense Codes are all standardized so a 03 is always a media error, but the other codes can be different. So usually the text that is added by the ARCserve engineers is correct, in this case Write Error, but it is possible it could mean something else. The way to be sure is to get the SCSI Reference Manual for the tape drive and look it up or ck with their tech support.

Now back to this specific case. The tape drive reported that it could not complete the Write command because it experienced an unrecoverable write failure. That means the drive tried to write, failed, retried, failed, etc. until its threshold was reached and it gave up on the Write command.

The cause could be a poor quality tape, contamination such as dust, environment such as heat. The owners manual for the drive gives the complete list of environmental factors that will case the drive to fail. Tapes are also an issue, the manufacture will have a specific list of recommended brands, I recommend sticking to that list.

 

[jarhoops]

Great, I have a technician coming to check the condition of the tape drives. We never reuse our backup media, and it's a recommended manufacturer.

Would it be possible to get this type of error if the SCSI BIOS settings were set incorrectly? Or would that generate another type of error? I'm just trying to think a step or two ahead in case we don't find anything wrong with the hardware. Thanks.

 

[vschumpy]

Unlikely, unless you have different settings specifically for the SCSI ID being used by the tape drive which you state works fine.

 

[davidmichel]

As vschumpy sates unlikely, but it is possible for the BIOS settings to cause the drive to report a write failure. The question is could it cause the drive to report this type of failure, and that is for the drive manufacture support to answer.