[Openstack] Recover Controller Node
giorgis at acmac.uoc.gr
Tue Oct 23 09:29:53 UTC 2018
the HDD on my OpenStack controller has failed and I would like to know
what is the best way to recover some basic controller functions assuming
that I can get some of the data from it.
Let me give you some information about the installation's background
and what I would like to do (ideally).
First of all let me say that this is an old OpenStack installation with
Icehouse on two bare metal nodes, one controller and one compute. Since
it was the controller's HDD that failed all my VMs are still up and
running but I have no way to perform tasks on them e.g. snapshot,
terminate, launch new etc.
So I am wondering what would be the best possible scenario and approach
in order to recover some functionality on them. What I would like
ideally to do is to be able to snapshot a couple of already running VMs
and then move the snapshots to a new installation that I am going to
deploy. Besides the snapshots I don't need anything else from the
The controller's HDD is not completely dead which means that I can
browse it and I can try to copy some files (e.g. older useful snapshots)
but clonning it to a new one wasn't successful (clonning worked but
couldn't boot). I can sure try to resurrect the DB from it and besides
that I would also need obviously all the "conf" files.
Assuming that I can get these from the failed HDD what would be my next
steps? Try on a different HDD to set the old environment from scratch
using the "saved" configurations? Would it be better if I moved all
required services, DB etc. to the compute node in order to perform the
maintenance on a single node? How can I do that? Should it be better if
I just go directly to a much newer version and how can I get the running
VMs from that? Any ideas what else I might need besides the DB and the
Looking forward for your answers!
More information about the Openstack