Faulty disk in volume group using Surestore FC10

[Cedric_8]

Hi,

I have following H/W config:
- A500 running HP-UX 11i with local boot/root disks
- FC10 disk system with 4 disks in it.

On the A500 we configured several vg (volume groups), each one containing 2 distincts pvg (physical volume group) so that we can mirror each logical volume attached to the volume group.

In vg10, one of the 2 disks is dead.

vg11 has 2 disks and only 1 lvol, not used so far.
We want to take one of the disks of vg11 and add it to vg10 to replace the faulty one.

So far we managed to do the following:
- in vg11, we used "Split a Mirror Copy Offline." on the only existing logical volume with sam. This created another logical volume and we were able to remove it and then mark the disk as unused.

Then, we tried to configure the new unused disk (using logical volume manager) into an existing volume group (with option "configure the disk into a pvg" to be sure to have a distinct pvg) but this failed without any message, it only said vgcreate and vgextend failed.

Maybe I have to do the split on each lvol in vg10 like the split I did on vg11. But then, how can I be sure it will keep the lvol that is on the right disk (not on the dead one). With sam you can't specify the disk you want to keep online (is it smart enough to decide for you ?).
I could use the command:

lvplit -g pvg101 /dev/vg10/lvol1
lvplit -g pvg101 /dev/vg10/lvol2
lvplit -g pvg101 /dev/vg10/lvol3

where pvg101 is the faulty disk and lvol1, lvol2, lvol3 are the existing logical volumes attached to vg10.

and then (?) do a:

vgreduce vg10 pvg101

So that vg10 only contains the sane disk.
Then afterwards I may be able to add to vg10 the unused sane disk taken from vg11 without any error.

What do you think about that ?
My constraints are:
- I cannot reboot the system
- I am not onsite

Another solution would be to use sam, mark the disk taken from vg11 as "ready to replace" so that it takes it out from the FC channel, take it off and replace the faulty disk with it.
But I'd prefer not to have to go onsite with the above solution if it's ok.

Thanks for your help, hope I've been clear enough.
Cedric.

PS: Here is what sam says when it tries to scan the hardware :

The Logical Volume Manager shows this device file, /dev/dsk/c4t5d0, with hardware path, 0/2/0/0.8.0.255.0.5.0, as belonging to Volume Group, vg10. SAM has determined that this hardware path is not currently active because either the device is disconnected, or this is not the primary path to a multiple path physical volume. This results in incorrect disk utilization information in the Disk Devices and Volume Group lists.

[Jannik · Accepted solution]

take a look at this link.
http://www2.itrc.hp.com/service/iv/node.do?node=prodITRC%2FWW_Start%2FN1%7C16%7C11

this is the LVM Bibel.

[Cedric_8]

It looks like I don't have access to this link. Maybe because I have an european login ?

[Robert-Jan Goossens]

http://www4.itrc.hp.com/service/iv/node.do?node=prodITRC%2FWW_Start%2FN1%7C16%7C11

[Cedric_8]

Damn, I've been RTFM'd ! :))
Thanks for the second link. It looks like I should find answers in that pdf.

[Cedric_8]

This document is really helpfull.
Still, there is a broken link in it that refers to another chapter:

3.For fibre channel disks perform the replace_dsk steps described in the section â How
to Replace Disks at Hosts with TachLite HBAsâ in the Fibre Channel chapter.

(links to P:/sw_recovery/source/Update/FibreChannel.pdf which obviously does not work).

Where can I find the whole document with all the chapters in it (you gave me link for chapter 16 regarding LVMs).

Thanks in ad

[Cedric_8]

Ok I found it:
http://www5.itrc.hp.com/service/iv/node.do?source=search&node=prodITRC%2FWW_Start%2FN1%7C16