Failure when recovering Spectrum Protect database on 8.1.21

[hbarnes]

Hi there,
I'm currently attempting to use the Starwinds VTL to perform a DR of an IBM Spectrum Protect/TSM 8.1.21 instance.
We're running the VTL in Local-Only mode, as the environment is air-gapped.
The TSM database has been written to the Virtual Tape, and appears totally fine. It has the standard 80-byte EBCDIC headers at the fore, then filemark, then the database itself. So i'm confident that the tape is _not_ the issue.
The VTL is running on a Windows Server 2019 VM, sharing over iSCSI to a RHEL 7 VM where we're attempting to recover the DB. The RHEL7 box can see the emulated 3584 library, it can see the 4 IBM drives and correctly assigns them with /dev/IBMchanger* + /dev/IBMtape* numbers.

When we attempt to recover the database using the `dsmserv restore db` command, it attempts to verify the label the volume and then proceeds to give us a series of errors.

Code: Select all

ANR8302E I/O error on drive DRV001 (/dev/IBMtape0) with volume NONE (OP=SETMODE, Error Number=22, CC=0, rc = 2683, KEY= 05, ASC=24, ASCQ= 00, SENSE=70.00.05.00.00.00.00.0A.00.00.00.00.24.00.00.00.00.00,
Description=An undetermined error has occurred)
....
ANR8353E 001: I/O error reading label of volume in drive DRV00 (/dev/IBMtape0)

Looking through the logs at the Starwinds VTL itself, I can see where it's attempted to read.

starwinds_log_error_1.zip

(590 Bytes) Downloaded 743 times

EDIT: Apologies did not see the post about not putting in truncated logs - Logs with VOLSER removed below.
The above logs still contain what I believe to be the pertinent part of the logs - All other functionality seems to work.

starwinds_full_logs_1.zip

(37.41 KiB) Downloaded 822 times

I think this means that something that TSM is doing that Starwinds does not like?
Is this a known issue, and are there any known workarounds?

Thanks

[yaroslav (staff)]

Welcome to StarWind Forum.
May I wonder if you have been using VTL for a while, or have you just downloaded it?
If the former statement applies, could you please let me know how long ago the issue started

[hbarnes]

Hi there,
I've only just started using it. I'm trialing the software to see if it's suitable for our use case.

Thanks

[hbarnes]

Update:
I've tried a few variations of this, and so far been unable to get this working.
I've tried defining the library as a VTL rather than a SCSI - No change.
I've tried setting DRIVEENCRYPTION=OFF on the devclass - No change.

All of these produce the same errors on TSM, and within the Server Log of StarWind.
"MODE_SENSE page 0x24, subpage 0x0 is not supported!" which my development team has told me refers to some sort of encryption check/sense that TSM does.

On the logs I see that the tapes are moved into the drive, the MODE_SENSE errors appear, and then the tape is ejected back into the storage slots. This takes about 0.4 seconds, start to finish.

Not sure if any of this information helps.
Thanks,
- Harvey

[yaroslav (staff)]

Harvey,

Could you please test if downgrading helps? I DMed you the previous build.

[hbarnes]

Harvey,

Could you please test if downgrading helps? I DMed you the previous build.

Hi there,
I've received no DM? My inbox is empty.

[yaroslav (staff)]

I will resend it sorry.

[hbarnes]

Hi Yaroslav, I installed the downgraded build and it appears to have bricked the VTL. I assume this is because the build you sent me does not support the IBM 3584 Emulation, which is preferred for TSM.
As such, I cannot test with this build.

Thanks,
- Harvey

[yaroslav (staff)]

Greetings thanks for your update. I will keep you posted on log investigation.

[yaroslav (staff)]

Please set the logging to LVL 3.
1. Stop StarWindService
2. Go to StarWind.cfg (see under C:\Program Files\StarWind Software\StarWind)
3. Copy it.
4. Edit StarWind.cfg.
5. Find <LogLevel value="1"/>
6. Set 1 to 3.
7. Save and exit editor.
8. Start StarWindService.

Reproduce the error and collect the new set of logs. Share them with me.
Once logs are collected, please undo the change mentioned above. You could roll back to the old *.cfg if needed.

[andrewmacleod]

Were you able to get this fixed? I tested it and encountered errors as well. I'm currently using Spectrum Protect with RHEL8 and the IBM tape driver.

[yaroslav (staff)]

Thanks for your post.
Please collect the logs in the way I described above.

[andrewmacleod]

logs.zip

(22.75 KiB) Downloaded 432 times

[yaroslav (staff)]

Thanks!
I have forwarded them to my colleagues. I will keep you posted.

[andrewmacleod]

I'm sorry, this is the right log... with debug level

The forum says it's too large to attach, so I put a link here

https://1drv.ms/u/s!Au-3Q1EAupF-gtYq3RM ... A?e=iXjtr5