Data Protector User Discussions
cancel

Medium marked Poor often

Highlighted
John Mathew_1
Absent Member.

Medium marked Poor often

Setup: Windows 2000 server (SP4 with all patches) as Cell manager With Patches- 69,61,53,79,77,67.
Device: Dell PV136T LTO-Ultrium Tape Library with 4 Drives
Clients: Windows 2000/NT server

I am getting lot of tapes marked as poor in the middle of any backup randomly in any of the 4 drives. Below are some messages i get.

[Major] From: BMA@corpomnint.corp.idt.com "Tape1" Time: 9/23/2004 7:43:47 PM
[90:182] Cannot forward segment. (Medium error.)

[Minor] From: BMA@corpomnint.corp.idt.com "Tape1" Time: 9/23/2004 7:43:47 PM
[90:65] scsi5:0:0:4C
Cannot append to medium (Medium error.)
[Major] From: BMA@corpgroup.corp.idt.com "GTape2" Time: 9/25/2004 1:17:59 PM
[90:51] scsi4:0:0:6C
Cannot write to device (Medium error.)

[Major] From: BMA@corpgroup.corp.idt.com "GTape2" Time: 9/25/2004 1:31:39 PM
[90:51] scsi4:0:0:6C
Cannot write to device ([121] The semaphore timeout period has expired. )

[Major] From: BMA@corpgroup.corp.idt.com "GTape2" Time: 9/25/2004 1:32:42 PM
[90:51] scsi4:0:0:6C
Cannot write to device (Details unknown.)

[Major] From: BMA@corpgroup.corp.idt.com "GTape2" Time: 9/25/2004 1:32:42 PM
[90:151] Apparently SCSI bus reset.



The same tapes when formatted and used with the standalone LTO drive, its not marked as poor.

Your help is appriciated
9 REPLIES
Leif Halvarsson
Absent Member.

Re: Medium marked Poor often

Hi,
Is the standalone drive connected to the same client and the same HBA as the library ?
Is perhaps the library connected to SAN ?

Check for any differences,
- Drivers.
- Drive firmware.
- Servicepacks and DP patches (if different clients.
- Also, check hardware (cabling, terminators etc.)

Kurt Beyers.
Absent Member.

Re: Medium marked Poor often

If the library is connected to the SAN, disable the Fibre Array Monitor in the HP Insight Management agents on the servers connected to the SAN.

Kurt
Stuart Whitby
Absent Member.

Re: Medium marked Poor often

Looks like a problem on your SAN.

What make are the drives? My guess is that they're not IBM, where these will automatically be set with persistent reserve.

Zone down your SAN as far as possible. I think that either resets or rewinds are going on from other systems behind the scenes, and you need to ensure that the only systems which can access those drives are systems which actually use the drives. Also, ensure that SCSI reserve is set up at a driver level - you may need to talk to Dell about how to do this. With this set up, nothing else should be able to access the drives while one system is using them. A SCSI reset will still break that reserve, but is likely to break the write at the same time, meaning you have a very noticeable issue.
A sysadmin should never cross his fingers in the hope commands will work. Makes for a lot of mistakes while typing.
John Mathew_1
Absent Member.

Re: Medium marked Poor often

Thank you Leif and Kurt,
I am not having a TRUE SAN environment, but the library is connected to a Fiber Switch and the Cell server is connected to the Fiber Switch also. There are couple clients also connected to the Fiber Switch, all these two clients and the cell server sees the Jukebox as locally connected SCSI devices.
Kurt, How to disable the Fibre Array Monitor in the HP Insight Management agents on the servers connected to the SAN?

thanks and regards
John
John Mathew_1
Absent Member.

Re: Medium marked Poor often

Thank you Stuart,
The drives are HP make though the whole library was bought under Dell brand name.
Looks like a problem on your SAN.
1.Please write me more on what is persistent reserve? and how to enable/disable it? I will talk to dell about this.
2. Only three systems are connected to this fiber switch including the cell server and all these three sees the Library as local SCSI devices.
3. Yes when these errors occur the backup fails.

regards
John

Scottie Cameron
Absent Member.

Re: Medium marked Poor often

John,

Any luck with this? I am having the same issue running DP 5.5, UNIX cell manager & a MSL6000 library... Have tried cleaning the drives several times and no luck. When I run the backup to a stand alone tape drive, it works, but in the library, I get these errors...

Scottie
scameron@roseville.ca.us
I said, I AM WORKING ON IT!
John R_3
Absent Member.

Re: Medium marked Poor often

I had that too on our SAN. I was adviced not to run sanmgr hostagent (on hp-ux do a ps -ef|grep diald) - if I recall correctly.

From the top of my head: I disabled hostagent with the command '/opt/sanmgr/hostagent/sbin/hostagent stop'

Not sure if there is something similar on a Windows box.

Not sure what this hostagent is supposed to do or whether it is important or not. No more wrecked LTO$$ tapes.

Kind regards, John
Scott McIntosh_
Absent Member.

Re: Medium marked Poor often

Shutting down the Storage Area Manager agent if you're actually using OVSAM is not likely the best answer. But SAN management tools can drop your SAN-based backups quite easily, so disabling it as a test is certainly warranted.

Thanks,
Scott
HP Support
John Mathew_1
Absent Member.

Re: Medium marked Poor often

Hi Scott,
The problem was with the HP Ultrium Drives. I replaced all 4 drives. Now there is no tape marked poor.
For details jmathew@idt.com
Thanks for all inputs

Regards
John
Top Contributors Last 30 Days