A mirroring question

[Mihails Nikitins]

Hi,

Imagine I have a volume group of 6 physical volumes, only one large logical volume, there are 3 mirrored pairs, mirroring is restricted by pvg.

Question.

When taking one disk away from mirroring for replacement, will the command break mirroring for all disk pairs?

lvreduce -m 0 /dev/vgxx/lvolx /dev/dsk/cxtxdx
vgreduce /dev/vgxx /dev/dsk/cxtxdx

Points in advance for your feedback!

BR,
Mihails

[Todd McDaniel_1]

you dont have to break the mirror to replace a disk...

just replace it then:

pvcreate /dev/dsk/cxtxdx (optional)
vgcfgrestore
vgchange -ay vgXX
vgsync

[Geoff Wild]

We had an issue on an internel disk recently - where EMS was reporting a failure - one of the file systems was gone - but the disk was still "available"....Here's what HP said:

Here is the information regarding disk replacement :

1) If the disk is completely "dead", such as if you run ioscan and status is
"no_hw" then you can hot swap the disk online.

2) However in circumstances where the disk has not fully failed please do one of
the following to avoid data corruption :

a) reduce mirror before replacing the disk
b) deactivate VG before replacing the disk
c) shutdown system to replace the disk

If the disk was not completely dead, I would suggest to deactivate VG if possible. There is a good kmine document on the steps I have mentioned above if you are interested :
KBRC00009115.

Rgds...Geoff

[Mark Grant · Accepted solution]

Good advice above but to specifically answer your question, yes.

You mirror on the logical volume level, not the disk level so you "lvreduce" the logical volume, you remove the mirror from entire logical volume.

[Mihails Nikitins]

Thank you for the replies!

The disk is not dead yet, just generating bad warnings, so I should execute LVM commands before pulling it out. I'm just unsure about handling one big logical volume that is based on more than one pair of physical volumes.

Mark, you've answered my question, thanks. If this is true let me I ask more.

Should I re-create mirror on the whole pvg after disk replacement?

lvextend -m 1 /dev/vgxx/lvolx /dev/dsk/cxt1dx /dev/dsk/cxt2dx /dev/dsk/cxt3dx

If so, should I bring offline all the disks in the "faulty" pvg first?

lvreduce -m 0 /dev/vgxx/lvolx /dev/dsk/cxt1dx /dev/dsk/cxt2dx /dev/dsk/cxt3dx

Points in advance!

[Todd McDaniel_1]

If the disk is not dead then yes you should definintely use all caution. You can only pull the disk if it is truly dead.

run diskinfo and pvdisplay on it to be sure as well...

Using LVM commands, lvreduce the mirror on all Lvols that live on that mirror.

Then replace the disk and then add the mirror back again.

[Mihails Nikitins]

Hi,

Just a summary.

1. taking one pv from mirroring will break the entire mirror.

lvreduce -m 0 /dev/vgxx/lvolx /dev/dsk/cxtxdx

There's no need to mention other disks in lvreduce command.

2. You need to mention all 'freed' physical when recreating the mirror.

Just identify what pvs are actually not used by commands 'lvdisplay -v' and 'pvdisplay -v'.

Then issue a command like
lvextend -m 1 /dev/vgxx/lvolx /dev/dsk/cxt1dx /dev/dsk/cxt2dx /dev/dsk/cxt3dx

3. I faced the problem described in DocumentID 4000009061.
That is becuase importing volume group removes
/etc/lvmpvg.

Just to keep archives better. :-)

BR,
Mihails