T10000C - Drive Hangs and Stops Communicating on the FC Bus and Needs to be Reset

Asset ID:	1-72-1379734.1
Update Date:	2017-09-20
Keywords:

Solution Type Problem Resolution Sure

Solution 1379734.1 : T10000C - Drive Hangs and Stops Communicating on the FC Bus and Needs to be Reset

Applies to:

StorageTek T10000C Tape Drive - Version Not Applicable and later
Information in this document applies to any platform.

Symptoms

Drive hangs and stops communicating on the FC or FICON interface and needs to be reset to recover.
T10000C DIA Channel Interface Hang issue.
PCI Bus Hang issue.
Stuck T10000C tapes
Input/output error reported at mount.
Unable to unload the drive.
4C06 SNO
IPL failure
The drive stops communicating over the channel interface.
Intermittent Hang issue.
The T10KC drive needs to be IPLed to recover.
May be affected by temperature..

Cause

Marginal PLL circuit between two devices on the TCM card.

Solution

This is not a data integrity problem. Not all non-reworked drives will fail.

If you believe you have a T10KC customer that is experiencing the drive hang condition, please open a collaboration to L2 tape for verification. L2 can help validate the issue and provide the next steps to alleviate the issue for the customer.

Process to determine whether or not a drive is a reworked drive:
On VOP use "Retrieve" then "View Drive Data" then select the "Version" tab.
The CO Hardware Features bits were changed when the rework went on.
The CO version will have a 1 in the third nibble (e.g 001031C3) if the rework is on the TCM card (zero otherwise).

Notes:
It is the third digit starting on the right hand side.
If this is a 1 the rework is installed.
If it is a 0 then the rework is not installed.

Additional Note:

We have seen drives fail even though the rework is shown to be on using the above check. The rework is delicate and is done under a microscope in our repair center. if you get a drive that has the symptoms for this problem it is best to force a drive dump and have it analyzed. Once it has been determined that the drive is taking a PCI Bus hang, a replacement drive needs to be ordered.

Part Numbers:
T10KC Drive (reworked for PCI fix) – 7039275 or higher . EC 0006621 that cut in this rework.

All drives that go through repair are reworked.

New build drives as of the last week of Sept 2011 have this rework. DMOD sn 1600 will have this rework. If your drive is below this DMOD number then you need to verify using the above process.

Library tray sn 1200 (SL8500 tray) or higher as new builds will also have the rework.

SCSI Toolbox version 9.0.0.0 and higher has a PCI bus test that can be used to check if the drive will fail. It writes several data patterns to a tape then tries to read it back.

It is during this readback that the test will fail if it is truely a PCI bus hang problem. The drive may hang and never come back at the point that it starts to read.

If you did not get a success message then the drive failed.

Note:

If you have a failure during the write phase of this test please ensure that the SL Protect feature is turned off in the drive as this is a known failure mode. If this is the case you

will get a failure on a write not on a read.

Attachments

This solution has no attachment