Replacing a faulty volume

naw_deepak · September 16, 2009, 1:51pm

Hi All,

I replace a sub-mirror of a mirrored volume by using below command:

#metadetach   d10(mirror)    d11(striped)
#metattach   d10    d13(other striped)

With above process I successfully replaced a faulty volume. I want to knoow if the process I followed is right. Or if not, what should be right step to replace it.

Thanks,
Deepak

TonyFullerMalv · September 16, 2009, 6:04pm

You could have used metareplace(1m), but that should be fine, run:

# metastat d10

and confirm that it is fully synchronised.

incredible · September 23, 2009, 11:24pm

Why didn't use
#metattach d10 d11
instead? Just asking..
You could also do
#metareplace -e d10 cxtxdxsx

naw_deepak · September 25, 2009, 6:33am

Hello Incredible,

#metareplace -e d10 cxtxdxsx

What it actually does? Because if we need to replace a slice by another than we must have to detach and attach the slice (striped).

Thanks,
deepak

incredible · September 25, 2009, 10:01am

You may wish to understand the definition from here
UNIX man pages : metareplace (1M)

pupp · September 25, 2009, 10:15am

all good stuff. however, please be cautious of the type of disk you are dealing with. metareplace works well with scsi disks but not so much with fcal. meaning, if you cfgadm a scsi disk, metareplace works. however, fcal (say a v880) you would need to break the mirror (deattach) and then after the disk replacement, attach it. also, this is just from what i've seen...