>
>> The disk layout on the older one is a 2 partition (/boot and /) with /
>> being a logical volume. Laying out what I think I need to do looks like
>> this :
>>
>> 1. yank old drive and cable up to newer system, expecting this 2nd drive
>> to now be /dev/sdb on the new box
>> 2. pvcreate /dev/sdb1
>> 3. vgextend VolGroup00 /dev/sdb1 (VolGroup00 is the volume group name on
>> both system, pretty much default installs in both cases)
>> 3. lvextend -l+100% /dev/VolGroup00
>> 4. resize2fs /
>>
>
> The biggest problems that I've had in similar situations are:
> 1) An entry in /etc/fstab that looks like
> LABEL=/boot boot ext3 defaults 1 2
> If the 'old' drive has a "/boot" lable, there may be issues. Either
> blow the old drive away or switch to a /dev/sda1 based fstab for the new
> machine
> 2) vg names the same on the 2 drives. Either rename them before moving
> the drive, or blow them away first.
>
>
>
I booted off a rescue disk and wiped the old drive first, so had no
problem getting it into the newer CentOS box and adding it there.
Unfortunately, it looks like the drive is now suffering errors (5
currently unreadable (pending) sectors, 3 offline uncorrectable sectors)
so now I need to undo what I just did, replace the drive (I have a new
one), and repeat. This is an older box, using IDE so I only have 4
(currently) occupied connections and I don't have room to add the new
drive while the failing one is still in there. Since I extended the
filesystem and logical volume to span both drives, but haven't written
any new significant data, the space I'm guessing is allocated for the
logical volume, but free. Does this checklist seem sane :
Original single disk was 80G; Added 2nd drive of 120G, so to undo I
should :
# reboot off CentOS-Live CD
# resize2fs /dev/mapper/VolGroup00-LogVol00 65G (ie. less than the 80G
on the first good drive)
# lvreduce -L 65G /dev/mapper/VolGroup00-LogVol00
(reboot and make sure filesystem is intact and nothing is sitting on
/dev/hdd1 with pvs )
# vgreduce VolGroup00 /dev/hdd1 (to remove drive from volume group)
# replace failing drive /dev/hdd1
(then I'll be back to where I started this exercise albeit with a new
good 2nd drive and can add it back in)
Smell sane ?
Thanks in Advance,
Jerry Queirolo
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
