Missing alternate link

[hamlite42]

Hi -
I have an N4000 running HPUX 11.0 that is connected by fibre to an IBM 2105 (Shark) system. The conection to the 2105 system is thru a San Data Gateway, wich converts fibre to SCSI. We replaced a bad SCSI cable per IBM from the SDGW to the 2105 system about 2 weeks ago. With 3 of the 5 volume groups created on this system I'm not seeing the alternate links anymore. When I do a strings on the /etc/lvmtab I see the correct volume groups with alternate links. Should I run a vgextend on each of the volume groups missing the alternate link?

Thank you.

Paul

[Animesh Chakraborty]

Hi,
Do a ioscan -fnC disk first.
Check whether you can see both tha path.


Regards
Animesh

[hamlite42]

Ran the ioscan -fnC disk & I'm missing the 3 drives.

[Michael Tully]

Hi,

You can't run a vgextend if you don't know what the device files are for those disks.
It could be that the drives just cannot be seen.

# ioscan -fn >/tmp/ioscan.out1
# cd /dev
# insf -e
# ioscan -fn >/tmp/ioscan.out2
Note the difference
# diff /tmp/ioscan.out1 /tmp/ioscan.out2

Sometimes with FC disk devices do not automatically get created if there is a change done.

HTH
-Michael

[Andre Da Silva Dias]

If you are able to disable the VG's, and if you haven't got BCV's of those disks visible on the box, you could:

- umount FS/disable VG's
- export VG's with shared option (vgexport -s -m vgXX.mapfile /dev/vgXX)
- Create VG's directories and group file
- import VG's with share option (vgimport -s -m vgXX.mapfile /dev/vgXX)

It will import the VG's with all alternate paths.

[Mark van Hassel]

Paul,

Just a "vgchange -a y" might help to update the running kernel with the /etc/lvmtab

[Sanjay_6]

Hi Paul,

Replacing a mere SCSI cable should not have spoiled you alternate link until you somehow changed the scsi path for the alternate link. Have you spoken to hp and ibm about this. A SCSI cable is merely a connection media with no active component or anything. Just do a little digging, and you may find you have overlooked a very simple thing.

hope this helps.

Regds

[Ian Dennison_1]

If they plugged the new cable back into another port, possibly the LUN has changed? Any new Drives present with a different "c[nn]" number?

Cheers, Ian

[hamlite42]

Thank for all the information! After running the ioscan on one system successfully, I tried it on another N server & the command is just hung up. The error message in dmesg is:

I/O hardware probe timed out
Last successful probe at path 0/2/0/0.8.0.50.0.3
I/O hardware probe timed out
Last successful probe at path 0/2/0/0.8.0.50.0.3
I/O hardware probe timed out
Last successful probe at path 0/2/0/0.8.0.50.0.1.0

I've check & restarted the San Data Gateway, but this did not fix the issue.
Thanks!

[hamlite42]

Hi - I tried the vgchange -a y & it was successfull on one system but on another N server the command is hanging. I'm thinking I need to reboot the system to clear the vgchange command. Any ideas?
Thanks.