- This topic has 3 replies, 2 voices, and was last updated 11 years, 4 months ago by
.
-
Topic
-
Hi. I’m running HA Cloverleaf with 2 Redhat nodes in a cluster on VMWare. I’m interested in what others with a similar setup do as far as backups are concerned. At this time, I’m just tarring the /hci directory. I’m trying to determine the best way to create some sort of bootable backup or if a routine clone of the nodes would be the best way. The way we setup our cluster, with a SAN in the middle housing the /hci directory, we can not clone an active node so a simple clone isn’t so simple. Thanks,
Mike
-
Replies
-
-
Are you set up with LVM? Using LVM, you can do a snapshot of the filesystem and then release that snap when the backup is completed. Snaps happen quickly enough that is you have a 2 minute a day maintenance window, you can do a snap with the engine down. You could also set up a rolling backup with each site and multiple snap, release, on the CL filesystem.
The other option is to snap it live. Easy enough to recreate the site database…
-
-
Nope. I usually stay “within the machine” as much as I can.
LVM is an abstraction of disk space between the physical disk presentation, and the file system (ext4, xfs, etc).
LVM consumes physical disk into Volume Groups, and you create Logical Volumes on those groups. You build your filesystems on those logical volumes. You can then resize those logical volumes on the fly, or snapshot the logical volume and mount the snapshot, and dismiss the snapshot when you are done.
I usually split up my physical disks into Logical Volumes that leave room in the Volume Group. (say 1 TB of raw disk… I allocate 500GB to filesystems)
If you then need more room in your /tmp or /opt/SMAT, you just resize it a bit larger on the fly. I don’t know anyone who can figure out exactly how much room they need for a filesystem 3 years before they need it.
-
- The forum ‘Cloverleaf’ is closed to new topics and replies.
