I had managed to splitbrain my lock files, and then screwed them up trying to heal them
manually.
I wound up backing up my hosted engine disk, destroying the hosted engine storage volume,
recreating the gluster volume and re-running setup to solve the lock issue, and copying
over my hosted engine backup instead of reinstalling it all. Re-added the host nodes, and
off it went. :)
Probably don't recommend that as a good thing, but it was my dev/testing cluster, so
it can be done without reinstalling from scratch.
-Darrell
On Aug 20, 2014, at 6:41 AM, Joop <jvdwege(a)xs4all.nl> wrote:
On 19-8-2014 17:08, Darrell Budic wrote:
> Through a series of creative experiments, I've managed to corrupt the lock files
for my self-hosted engine's storage pool. Is there any way to destroy and recreate the
pool without reinstalling the self-hosted engine on all systems (after destroying it
entirely)?
>
>
Don't know if my situation reflects yours but I had a problem with
hosted-engine too. I had a way newer version of kvm/libvirt on host01
and decided to yum remove it. Sofar no problem but it also removed
hosted-engine and then i had a problem. In the end I solved it by just
reinstalling hosted-engine and running engine-deploy on host01 and
answering the question if this was a reinstall and pointing to the
anwser file on host02. Once it was done all came up nicely.
So it seems possible to maybe do this in a roundrobin way and
update/reinstall your hosts.
Joop
_______________________________________________
Users mailing list
Users(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/users