Identifying RHEL disk other than by UUID in /etc/fstab for Disaster Recovery from another site

Hanne Winter asked:

For our disaster recovery setup there is a set of disk which are replicated from our production site to the disaster site. On the production site the RHEL servers have /boot filesystem setup on disks specified via UUID in the /etc/fstab .

However when the DR counterpart server at the disaster site is booted up with the replicated root disk from the disaster site storage array instead, then the UUIDs for the disk with the production data are different so that the system cannot find the /boot device.

So I am hoping to find some thoughts on the best practice for setting up /etc/fstab so they will create consistently even where the UUID for the physical devices may change.

The reason that UUID is chosen to reference disk in /etc/fstab is due to these being SAN attached disk. So if target path to disk changes or something else, then we need to make certain that /boot FS will always be found via something unique.

Filesystems on the logical volumes are referenced using LV path instead.

My answer:


For /boot you should be able to label the filesystem and then refer to it by label. Indeed, by default RHEL already labels /boot as /boot.

e2label /dev/disk/by-uuid/**** /boot

will change the disk’s label to /boot.

You can then refer to it in /etc/fstab:

LABEL=/boot       /boot           ext3            rw,relatime        0 2

LVM volumes should just be referred to by their VG/LV as you are already doing.


View the full question and answer on Server Fault.

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.