making the HP VME Console HA

[mikejdunphy]

Hello,
So were working thru the trial and I already killed the vm essentials console by having the node it was running on fill up on disk,
I was testing uploading virtual images. I created a large virtual image NFS share to use but apparently it still uses the local disk
for caching or temporary space. In any event when the disk filled up it moved it to node 3 and of course it cant run there since
this is the first vm built and it was done only on node 1. So it was halted and dead. So I figured out how to get it off node 3 and start back up on node 1after I had allocated more disk to node 1 of course.
So here is my question what is the easiest way to make this HA and migrate the vm to a 10TB GFS2 datastore that I have mounted
on all 3 nodes so I dont have to deal with this COW image growing and getting too big ?

[mikejdunphy]

will this work ? I want to be careful because were in a trial and I dont want to blow it up and start
all over with a new console. In fact if we can fix the root problem which is the provisioning in another
post that is filling up the local disk I would probably just leave it till after the trial and if we decide
to buy it and use it fix it then

# On node1 (current host of hp-vmemanager)

virsh shutdown hp-vmemanager
# Wait until it's fully off
virsh domstate hp-vmemanager

# Move the qcow2 file
# given that the gfs2 mount is /mnt/97fd47ae-aa9a-4fd5-a16e-2af49d9e9fe7
mv /var/morpheus/kvm/local/hpe-vm-essentials-8.0.6-3.qcow2 /mnt/97fd47ae-aa9a-4fd5-a16e-2af49d9e9fe7/

#probably should create a directory structure tho
mkdir /mnt/97fd47ae-aa9a-4fd5-a16e-2af49d9e9fe7/hp-vmemanager
mv /var/morpheus/kvm/local/hpe-vm-essentials-8.0.6-3.qcow2 /mnt/97fd47ae-aa9a-4fd5-a16e-2af49d9e9fe7/hp-vmemanager

# Edit the domain to point to new path
virsh edit hp-vmemanager
# Change the <source file=...> under <disk> to:
# /mnt/97fd47ae-aa9a-4fd5-a16e-2af49d9e9fe7/hp-vmemanager/hpe-vm-essentials-8.0.6-3.qcow2

# Start the VM
virsh start hp-vmemanager

[Arnout_Verbeken · Accepted solution]

Seems that you did not add the GFS2 to your cluster, and if you did that already, it's a best practice to make the local datastore inactive so VM's don't land on the local datastore. (by default, when creating a new VM, VME will choose the datastore to land it on if you have multiple datastores).
To add GFS2 to the cluster: cluster - storage - add - gfs2

The VME manager itself is indeed installed on the local datastore.  Once you have configured your GFS2 datastore, you can move the VM via the "reconfigure" option on the VM and just select the GFS2 datastore for the disk (instead of the local datastore it will be on).
VME will do now a "Storage Vmotion" of your VM.

So there is no need whatsoever to ever do anything in virsh yourself (except maybe to just take a look and/or to troubleshoot)

Hope this answer is sufficient.  Otherwise, drop a message here.

[mikejdunphy]

yes it has been added, In fact it was created with vme based on input from someone else from HP on this forumn I will try reconfigure and move it off. Thanks

[mikejdunphy]

awesome it worked great thanks

[Thaufique_Mod]

That's great!

We are delighted to hear that your concern has been addressed.