Changing disk device numbering

[John Tyler]

Hello!

We have 2 N-Class running 11.0 connected to VA7100 in FC_AL mode. We replaced the hub with a switch and the device numbering changed and is different on each server.

The fiber adapters are on the same i/o slots on each server

I need to setup MC/SG. So i want to keep the device numbering the same.

Example: The disks are seen are c2t2d0-c2t2d6 on one server. The same disks are seen on c3t2d0-c3t2d6 on the other server.

How do i change the device numbering on one server to match the numbering on the other.

TIA

[Armin Feller]

Hi TIA,

here a little cookbook which described HOW TO CHANGE THE INSTANCE NUMBER:

--------------------------
1. Extract a configuration template from the current ioscan output.
Execute the following command:

# ioscan -f | grep -e INTERFACE -e DEVICE | grep -v target | awk '{print $3, $1, $2}' > /infile

2. Edit /infile and change the ext_bus and lan instances as desired.
No class is allowed to get more than one line for the same instance!

3. Bring down the system gracefully to run level 1.

# init 1

4. Apply the ioconfig change:

# /sbin/ioinit -f /infile -r

The system will reboot immediately if the change is successful.
Warnings like 'Input is identical to kernel' can be ignored.

If unsuccessful, the most likely error to happen is:
"ioinit: Instance number X already exists for class XXX"

The problem is that your desired instance assignment conflicts with
an existing instance number. If that instance is bound to hardware
that is no longer visible in ioscan, then you are in trouble and
need to perform the Procedure II or III.

5. Once the system reboots, verify that all the instance numbers
were changed as expected. It may be necessary to re-import volume
groups to ensure that /etc/lvmtab contains the correct
entries. The lan configuration may need to be changed also.
--------------------------

Best regards,
Armin

[Thayanidhi]

Hi,

Whenever you change ID, hp-ux creates new controller instance! (c?t2d0)

You may try #rmsf both c2t2dx and c3t2dx
(use "-H" option and specify h/w path)
and "#insf -e"

This is what I do. But I am sure about your environment.

TT

[Thayanidhi]

I am sorry I missed "NOT" before 'sure' in my post.

TT

[John Poff]

Hi,

You can do it, but it isn't real easy to do. Why not just leave it like it is? I have several MC/SG clusters where the shared disks have different device numbers for the shared disks.

This reminds me of a funny story. We went through a similar exercise about 4 1/2 years ago when we setup our first HP systems. We had two V-class boxes and we were setting them up in a MC/SG cluster. We had a really, really sharp consultant (who works for HP now) who was doing the work for us. He was trying to get the device numbers the same for the shared disks between both nodes. Our man decided that if he could just remove one card from one V-class box, reboot and do some magic [rmsf, insf, etc.], and the put the card back in he would have what he wanted. He shutdown one V box, pulled the card, did his magic, got the device numbers the same, put the card back in and rebooted. Only problem, the V-class wouldn't boot. Since we had done the work ourselves without calling HP, we were potentially in some deep doo-doo. He gave it several tries and couldn't get it to boot. We decided to bite the bullet and call HP, but before we did that we decided to put the box back together, which meant that he had to put back on the cover for the cards, which was a long metal piece that fit really tight. Our guy was stretched out on the floor, trying to get it in place, and it wouldn't go. Finally, getting frustrated with the whole affair, he looked up at me with a real look of rage on his face and pounded the piece with his fist. Now, this guy is about 6'6" and on the high end of 200 lbs., so he hit it with a lot of force! It popped right into place. We put the skins back on, and decided to give it just one more try at booting before calling HP. Surprise! It booted up and ran like a champ. And we never, ever did anything like that again. So, now you see why I'm not so worried any more about having the disk device numbers match up between two nodes.

JP

[Bill McNAMARA_1 · Accepted solution]

you're entering into a world of pain...
The easiest thing to do, is ensure that the hardware is IDENTICAL when installing, remember it for the next time!

Other than that, yes, as mentions you need to recreate your io maps which is messy.

see:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x8f567b8d1de3d5118ff40090279cd0f9,00.html


and for a html cookbook
http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xb5968cc5e03fd6118fff0090279cd0f9,00.html

[Simon Hargrave]

You don't need the device names/numbers the same on each node. Simply create your VGs/LVs on a "primary" node, vgexport -s -p -v for each volume group, then vgimport on other cluster node(s). LVM will find the correct disks based on the PVID, and create the volume groups accordingly. Don't worry about matching up Instance numbers etc, that's headache territory. And you don't want a headache like that just before Xmas now, do you?!