Note: This article was originally written for SUSE OpenStack Cloud 6, and updated for SUSE OpenStack Cloud 7. It may not apply to later SUSE OpenStack Cloud releases.
If you’re using SUSE OpenStack
Cloud, you may
want to erase and reinstall your cloud deployment a few times during
the testing or proof-of-concept phase. You may also want to experiment
with a few permutations of Crowbar network configurations. SUSE’s
(otherwise excellent) Deployment
suggests that the only way to change your Crowbar settings, after
install-suse-cloud has been run, is to reinstall your entire admin
That isn’t really true if you know what you’re doing.
You may be thinking that you could just use
revert to your last Btrfs snapshot
created before you ran
install-suse-cloud. After all, running
crowbar, like any other YaST module, automatically creates a
before-and-after Btrfs snapshot of your root filesystem and all its
subvolumes. So, reboot machine, select pre-
snapshot, complete boot, run
snapper rollback, done. Right?
Well, not quite. If you
followed the Deployment Guide closely,
you will have removed your Btrfs subvolume for the
and replaced it with a separate, XFS-formatted partition. That means
it is excluded from all
snapper Btrfs snapshots, and thus, no
rollback for you for that directory. Which, of course, Crowbar uses
So, here is your checklist for resetting your admin node to a
Reboot your admin node.
In the SLES boot menu, select an appropriate snapshot taken immediately prior to running
Boot into your snapshot.
After rebooting, delete the following and directories:
- all subdirectories under
Then, you can reconfigure Crowbar (
yast2 crowbar), run
install-suse-cloud, and reboot your OpenStack nodes. They should be
discovered anew, and you’re then able to redeploy your OpenStack
barclamps to them.
This article originally appeared on the
hastexo.com website (now defunct).