MSL5026 + DP5.1 drive going offline

[Remon van Gijn]

We have a running issue with a MSL5026 Library 2x DLT40/80 drives, where DP puts the tapedrive 1 offline after finishing with it.(is scsi channel where drive and library are connected) The scsi board somehow notices the Drive offline as a Hardware failure causing a snmp trap & alert to our management systems. (storage agents 7.0.00) After DP is done and closes the sessions the drive comes back online, and all is okey again. Except for us having to deal with a fake hw major every time.

Anybody know of this or has tips?
(TT&L gives same info right after backup, after +-20 minutes all is okey again also in TT&L)
-Question posted as well in HPOV-DP forum.

Thxs for the mind minutes.

[Kurt Beyers.]

Is your firmware up-to-date? And the same for the DP patches.

Kurt

[Remon van Gijn]

yes both are the latest.

Firmware Revision 4.23
Hardware Revision 0
Boot Version 3
Library SCSI Id 0
Library Mode Random
Library Status Idle
Hours Since Power On 170


Drive 0 Status (Summary)
Drive Type DLT 8000
SCSI ID 1
Hardware Error? NO
Compression On? NO
Tape Format No tape
Tape Motion --
Write Protected? --


Drive 1 Status (Summary)
Drive Type DLT 8000
SCSI ID 2
Hardware Error? NO
Compression On? NO
Tape Format Unknown
Tape Motion Calibrating
Write Protected? NO

We downloaded the last HP tape drivers from the website.

[Gilles Cadrin]

First:

If running WINDOWS flavor .. You shouldn't be using tape drivers with DataProtector .. It talks directly to the scsi controler level ...

Also If using Win2k or Win2k3 The library shouldn't be taken in charge by the removable Storage service .. Please look at the Release note ...

So your device path should be scsi?:?:?:? instead of Tape?:?:?:?

Second You should update the your compaq insight manager ... A tape drive shouldn't be pull for a status during backup operation .. It needs an exclusive access ... They are known issues with older version of CIM ( Compaq Insight Manager ) On The HPUX side, they have the EMS software that basically does the same thing and the POLL_INTERVALL as to be uncommented out and set to 0 in he dm_stape.cfg file ...


DP Put the drives offline in order to eject the tape back to the slot ...

The reasoning is the not all tape drives adopted the standard of eject included an offline first ... A Goog Example would be an STK drive ...

Gilles

[David Ruska]

> Second You should update the your compaq insight manager ... A tape drive shouldn't be pull for a status during backup operation .. It needs an exclusive access ... They are known issues with older version of CIM ( Compaq Insight Manager ) On The HPUX side, they have the EMS software that basically does the same thing and the POLL_INTERVALL as to be uncommented out and set to 0 in he dm_stape.cfg file ...

The problem was actually with Compaq Insight Management Agents (CIMA?), and not CIM itself. The problem was with the 6.40 (and 6.20/30 with softpaqs). The problem was rectified in the Insight Management Agents version 7.0.0.0. These agents are in the Proliant Support Pack 7.00A, but they are NOT in the HP Insight Manager 7.0 SP2.3 bundle (yes I know, very confusing).

Remon, It sounds like you are running the Insight Management Agents version 7.0.0.0, is that correct?

[David Ruska]

FYI, you can check the version of management agents by going to windows control panel, and opening up "HP Management Agents".

More than a few folks have been confused by their version of Insight Manager (which is independent of the agents).

[Remon van Gijn]

Thxs for the tips so far !
The facts :

-We are running windows 'flavour'
-We do not use the tape drivers (are disabled)
-The drives are adressed using scsi?:?:?:?
-The Remote storage service is disabled
-The changer is steared by a HP driver in DP called "Changer:?:?:?:?" ?=numbers offcourse
-We are running all the last managers included in SP 7.0.0A with the IM agents 7.0.0.0

(the whole installation & hardware is 30 days old, so all all is brand new)

Meanwhile we did get a notify from HP support that the agnets of 7.0.0.0 also might contain an error related to the scsi polling. A new release was scheduled in 2 months. Advice included the sugjestion to disable the storage agents.
I think not smart as the other parts are not monitired either then, and windows service manager reported dependencies as well.

Also survey service was asked to disable.

Anybody other tips or have the same trouble ?

[D Wong]

Hi Remon:

I have a question about what exactly happens when DP puts tapedrive 1 offline.

1) Does the backup job fail when the drive is in this state?

2) If there is no backup job failures, is the problem that the network monitoring station for SNMP traps continues to receive hardware failure errors for the tape drive until DP is done with the backup job?

3) Does disabling only the Insite Manager monitoring of the SCSI devices solve this problem? (only disable the SCSI monitoring agent part)

Thanks,
Dave W.

[Remon van Gijn]

Hi Dave,

1) No the backups work 100% Okey. no failures etc.

2)Yes the monitoring keeps sending the hardware errors untill after the backups are done, and then the drive is returned according to the log and also eventlog in w2k from Offline to OK (status 5 to 2 according to the agents)

3)yesterdag I removed the SCSI Information agents and Storage Remote Alerter from the HP Managemrnt agents control in control pannel. Also the survey service was turned off by me and today indeed I had no errors anymore.

BUT:... What does it still report now? Disk failures, orger errors? I think we are blind to thses now. Is this correct ?

I will turn all on again besides the Storage remote alert, and do another backup.

[Gilles Cadrin]

By The way .. The changer should also be disabled .. It should say scsix:x:x:x and not changer


Gilles