LTO3 NO_HW on HPUX 11.31 after drive replacement

[wurzul]

Hi,

We have an MSL6060 tape library with 4 LTO3 drives connected to our SAN via 2 HP StorageWorks e1200-320 4G routers.

These drives are then multipathed to a number of HP-UX servers.

Recently one of these tape devices failed was replaced by an HP engineer.

Problem is The replaced drive is being picked up on our 11.23 servers but not on the 11.31 servers.

I've asked our SAN team to check zoning and all seems to be OK. I dont think the e1200 are at fault because the 11.23 servers have identified the device OK.

I removed the special files for this tape device using rmsf -H hw_path and then rescanned hardware using ioscan

No luck, below is the current status.

# ioscan -fnCtape
Class I H/W Path Driver S/W State H/W Type Description
==================================================================
tape 0 0/3/0/0/0/0.225.0.255.0.4.0 stape CLAIMED DEVICE HP Ultrium 3-SCSI
/dev/rmt/0m /dev/rmt/c3t4d0BEST
/dev/rmt/0mb /dev/rmt/c3t4d0BESTb
/dev/rmt/0mn /dev/rmt/c3t4d0BESTn
/dev/rmt/0mnb /dev/rmt/c3t4d0BESTnb
tape 1 0/3/0/0/0/0.225.0.255.0.4.1 stape CLAIMED DEVICE HP Ultrium 3-SCSI
/dev/rmt/1m /dev/rmt/c3t4d1BEST
/dev/rmt/1mb /dev/rmt/c3t4d1BESTb
/dev/rmt/1mn /dev/rmt/c3t4d1BESTn
/dev/rmt/1mnb /dev/rmt/c3t4d1BESTnb
tape 3 0/3/0/0/0/1.149.1.255.14.4.2 stape CLAIMED DEVICE HP Ultrium 3-SCSI
/dev/rmt/3m /dev/rmt/c4t4d2BEST
/dev/rmt/3mb /dev/rmt/c4t4d2BESTb
/dev/rmt/3mn /dev/rmt/c4t4d2BESTn
/dev/rmt/3mnb /dev/rmt/c4t4d2BESTnb

Is there any additional steps that need to be done so that the 11.31 servers can see this hw again?

Many thanks

[Alzhy · Accepted solution]

you would have to do a scsimgr gyration. Check your syslogs/dmesg.

[Marino Meloni_1]

look at this thread:

http://h30499.www3.hp.com/t5/Tape-Libraries-and-Drives/Replacing-Tape-Drive-Configuration/m-p/4729801#M26601

[Ralf Loehmann_2]

Hello Wurzul,

In HP-UX the device file for a Device in the SAN uses a combination of the WWN and the Serial number of the Device (inquiry page 83). This means if one or the other is changed the device file identification is changed. The idea behind this is HP-UX is always on and you want to change for example your tape library to a different switch in the SAN you just change zoning and than move the device to the new ports. On HP-UX you might not need to change something in that case. Bad thing on a repair the one or the other is changing and you need to fix that.
Here is a better thread which explains what to do and also there is background information about the command in the whitepaper link. Also it works to reboot, but that is something an HP-UX customer does not like.
http://h30499.www3.hp.com/t5/System-Administration/scsimgr-to-replace-disk-and-tapes/m-p/4501840#M364841

Regards,
Ralf.

[wurzul]

On both servers 11.31 servers I have run

/usr/bin/scsimgr replace_wwid -D

this has fixed the issue and the devices now shows as claimed again

Many Thanks to all !