Cisco Switch Migration affecting Linux RHEL5

[joe_91]

SAN team is migrating Cisco Switches. On HP-UX we need to do vxexport and vgimport as the c#t#d# would change. We are prepared for that. On the Linux RHEL boxes we are not sure what to expect? Anyone has any ideas what we need to do before and after the migration other than verify the dual PATHS to SAN??

Thanks

Joe.

[Tim Nelson]

first thought is if you have multipath configured with alias names then you will see some new paths (sda...sdxy etc) registered to those WWIDs and some faulty paths.

easiest way to remove the faulty is to reboot.

[Tim Nelson]

oh, and if you use the /dev/sdX devices files in any of your configs, e.g. lvm or fstab then you will need to fix them all..

print out your mappings so you can fix, or migrate to using multipath names.

[joe_91]

Please could you expand on that?

Thanks

Joe.

[Alzhy · Accepted solution]

No need to vgexport Joe. Assuming all the zones are recreated properly so that your SAN disks are still presented after the migration:

Shutdown RHEL server
Let them do their SAN Gyrations
Start RHEL Server

You should have all your LVM and other disks (ASM or ASMLib, etc)still intact and working fine. LVM PVs are agnostic to /dev/sdNN (or /dev/dm-NN) device name changes.

If you use custom firednly names in /etc/multipath.conf -- then the better as it will allow you to quickly assess if all your disks are back.

With Luv.

Alzhy Le'Cruel