10-30-2013 03:01 PM
I have an LVM version 2.1 volume group that was created months ago using the newer disk device file names, so when vg05 was created, disk53, disk58 and disk62 were all included.
Yesterday, we lost a fibre channel switch and had to replace it. The network guys inadvertantly changed the high level domain number when reconfiguring the switch. So, in hardware addresses that had been 0/5/1/0.10.63.0.0.0.1 the number "10" became a "1", the new addresses are like 0/5/1/0.1.63.0.0.0.1, etc...
Is there anything I need to do to vg05 in order for it to recognize those new hardware addresses? I am more accustomed to lvm 1.0 where I would vgextend and have alternate paths, but I'm assuming that's all just going to work by itself now with lvm 2.1?
Thanks in advance,
Solved! Go to Solution.
10-31-2013 06:38 AM
It's not so much the LVM 2.1 VG that will take care of this as it is the new I/O stack on HP-UX 11.31.
With 11.31 the OS will see and detect disks that point to the same physical disk/LUN and act accordingly.
If you run an 'ioscan -m dsf' you can see a map of the persistent DSF to the legacy DSF. An 'ioscan -m lun' will show you a bit of a different view.
As long as you can still see the disks through the new switch you should be OK.