If there is no enough free physical extents (PV) in the volume group, we have to add a new disk (as /dev/sdd) and its size need to be large than the replaced disk. In the moment, we also need to know which disk (as /dev/sdb1) need to be replaced or view /var/log/messages to find out the bad disk.
After done it, we have to make this new disk to be available with LVM by using pvcreate /dev/sdd command
To verify the status of the new physical volume by using pvdisplay /dev/sdd command
To realize which volume group the bad or old disk ( /dev/sdb1) belongs to by using pvdisplay /dev/sdb1 command
To add the new disk ( /dev/sdd) to the volume group drbdVG00 by using vgextend drbdVG00 /dev/sdd command
To realize which logical volume the volume group drbdVG00 belong to by using lvdisplay command
To move physical extents in active volume group drbdVG00 from /dev/sdb1 to /dev/sdd by using pvmove –n drbdVG00 /dev/sdb1 /dev/sdd command
To remove bad or old disk from the volume group by using vgreduce drbdVG00 /dev/sdb1 command
Finally, please shutdown the system next to remove the bad or old disk (# 1) and boot it. The data still exist in LVM disk as before.
Reference :
(# 1) When remove the bad or old SCSI disk but never change disk order (that the new disk ID also is the same as before)
all disk order will be changed but LVM function is normal.
When remove the bad or old SCSI disk and change disk order (that new disk ID is the same as the old one)
the bad or old disk just is replaced with the new disk and LVM function is normal as well.
