System Administration
cancel
Showing results for 
Search instead for 
Did you mean: 

Please Help - WWN change for LUN presented from enterprise SAN...

Greg Taylor
Advisor

Please Help - WWN change for LUN presented from enterprise SAN...

We have an old rp2450 running HP-UX 11.11 that has a LUN presented to it from our enterprise SAN. The SAN team is going to make a path change that will cause the WWN on their end to change, though the actual data/storage on the SAN will not change.

 

Our plan is to bring the server down, they'll make their change, and then we'll bring the server back up.

 

My question is, when then server comes back up, will it automatically see the LUN and recognize it as it was, or will I have to change things on the server in order to get it working again?

 

Here are some other considerations...

 

• The LUN is being used as the only disk in a volume group.

• The volume group contains logical volumes for file systems (not related to the OS) and raw devices used by Sybase.

• And it contains some swap. I think the admin before me needed to double swap for some reason and there was not any room on the internal disks so he used the LUN from the SAN. (There is 4 GB of swap on the internal disks and 4 GB on the SAN).

 

The server is going away in a few months so I'm not concerned with the ramifications of having swap on the SAN, except where the WWN change may be impacted by it.

 

What, if anything, will I have to do on the server once I bring it back up (after the WWN on the SAN side changes)?

 

 Thank you in advance for your help and advice!

 

 

 

 

4 REPLIES
Patrick Wallek
Honored Contributor

Re: Please Help - WWN change for LUN presented from enterprise SAN...

Before you shut down the system, I would do the following:

 

1 - Run an 'ioscan -kfnC disk' so you have the disk information as it is currently.

2 - Run 'vgdisplay -v' for the VG so you have the VG config.

3 - Run a 'vgexport -p -v -m /tmp/vg01.map /dev/vg01' (this assumes your vg is vg01, if it is not, use the correct name)  This will generate a map file that can be used to easily recreate the VG if it does not activate automatically due to the disk device file changing.

 

After the change and the system has been restarted:

 

1 -- Check to see if the VG was activated and your file system mounted.  If not:

     A. Run an 'ioscan -kfnC disk' to see what the new disk is, if it changed.

     B. If you have to recreate the VG, you can do the following:

          a.  Run 'vgexport /dev/vg01' to export the current VG

          b.  Make the /dev/vg01 directory (mkdir /dev/vg01)

          c.  Make the group file: mknod /dev/vg01/gropu c 64 0x010000

          d.  Import the VG -- vgimport -m /tmp/vg01.map -s -p /dev/vg01

          e.  Activate the VG -- vgchange -a y vg01

          f.  Mount your file systems -- mount -a

 

 

Greg Taylor
Advisor

Re: Please Help - WWN change for LUN presented from enterprise SAN...

Thanks Patrick,

 

So it looks like you think there is a chance I may need to export and then recreate the volume group via an import... That makes sense.

 

I hope the LUN just shows up and is recognized as it was before the change without intervention on my part. But if not, I really appreciate you pointing me in the right direction.

 

I wonder if the vg is not recognized, if the swap that is located there will cause any issues with it booting up. I assume not since there is still swap (with a higher priority) on the internal/boot disk. The swap missing from the LUN would be the same as commenting out the swap entry in /etc/fstab. There would likely be less swap than desirable, but it should still boot <fingers crossed>, and then I can work to get the vg recreated via an import if needed.

 

Thanks again!

 

Please let me know if you (or anyone else) thinks of anything else I should consider before we have to make the change this weekend.

 

Patrick Wallek
Honored Contributor

Re: Please Help - WWN change for LUN presented from enterprise SAN...

If the VG does not activate automatically, then you are correct in that your secondary swap will not be available.  Your system should still boot since your primary swap is OK.

 

Once you (re)import the VG you can do a 'swapon' on your secondary swap area to activate, or you could just reboot again to make sure everything is  mounted automatically as you expect.

Greg Taylor
Advisor

Re: Please Help - WWN change for LUN presented from enterprise SAN...

Patrick,
 
As suspected, upon reboot after the SAN changes were made the server did see the LUN as a new disk and therefore could not activate the volume group, etc. Though things went well (thanks to your help), there were a couple gotchas I thought I'd point out for the benefit of others in case they ever have to work through a similar situation...
 
1. Before making any changes or shutting down the server, also save a directory listing of the volume group device files (ie: /dev/vg01 or whatever your VG is named). After importing the volume group using the map file and new disk, the ownership of all raw LV device files will be root and in our case several had to be owned by Sybase in order for Sybase to use them. A directory listing would be helpful when restoring ownership on those files.
 
2. I modified the steps for recreating the volume group a little as follows:
a. Run 'vgexport /dev/vg01' to export the current VG
b. Make the /dev/vg01 directory (mkdir /dev/vg01)
* I added 'chown root:root /dev/vg01' so ownership would be as it was before the change
c. Make the group file: mknod /dev/vg01/group c 64 0x010000
d. Import the VG -- vgimport -m /tmp/vg01.map -s -p /dev/vg01
* I changed to 'vgimport -m /tmp/vg01.map /dev/vg01 /dev/dsk/c#t#d#' (I had to specify the new disk inorder to get it to work.)
e. Activate the VG -- 'vgchange -a y vg01'
f. Mount your file systems -- 'mount -a'
 
In my case I also had to run 'swapon -a' to activate the additional swap that is part of my volume group.
 
However, I did not specify the second disk as an alternate link in step 2d. So before I started my VG looked like this:
 
--- Physical volumes ---
PV Name /dev/dsk/c9t0d0
PV Name /dev/dsk/c11t0d0 Alternate Link
PV Status available
Total PE 31995
Free PE 7994
Autoswitch On
Proactive Polling On
 
and now it looks like this (it is missing an alternate link):
 
--- Physical volumes ---
PV Name /dev/dsk/c13t0d0
PV Status available
Total PE 31995
Free PE 7994
Autoswitch On
Proactive Polling On
 
Is there an easy way now for me to add /dev/dsk/c15t0d0 as an alternate link to the PV for this VG (without disrupting the system)?
 
Thanks again for all your help!
 
 --Greg