We have 2x Promox VE (1.5) boxes on which we are running a handful (say 15 each) lightly loaded VMs (all using OpenVZ, no KVM).
We are very happy with how this is running and now want to consider how we achieve quick recovery of VM's should one box fail.
Currently the boxes are in a cluster for management ease purposes.
Note that:
- There is enough grunt (and disk) available on each box to handle all the VM's being run from just 1 box
- There is only local storage on each box (no shared)
- We are happy with the VM's that boot being old in the case of an emergency
- We would like to have no more than say 5 minutes of startup of the VMs on the 2nd box should the other one fail
So obviously it is easy for us to script up rsync of /var/lib/vz/private/XXX over to the other box on a nightly basis, and I
guess similarly so for /etc/vz/conf/XXX.
Because we're not overly familiar with proxmox yet, the questions are:
- What else would need copying (if any)
- Is there any gotcha's we need to look out for doing it this way
- What would be the best way to start the VM's on the second box
In an ideal world, we'd receive notification of one of the boxes being offline, and simply logon to the second box and
run a script which would start all the copies of the other VMs (that were rsync'd the night before).
Thanks for any guidance.
Brent
We are very happy with how this is running and now want to consider how we achieve quick recovery of VM's should one box fail.
Currently the boxes are in a cluster for management ease purposes.
Note that:
- There is enough grunt (and disk) available on each box to handle all the VM's being run from just 1 box
- There is only local storage on each box (no shared)
- We are happy with the VM's that boot being old in the case of an emergency
- We would like to have no more than say 5 minutes of startup of the VMs on the 2nd box should the other one fail
So obviously it is easy for us to script up rsync of /var/lib/vz/private/XXX over to the other box on a nightly basis, and I
guess similarly so for /etc/vz/conf/XXX.
Because we're not overly familiar with proxmox yet, the questions are:
- What else would need copying (if any)
- Is there any gotcha's we need to look out for doing it this way
- What would be the best way to start the VM's on the second box
In an ideal world, we'd receive notification of one of the boxes being offline, and simply logon to the second box and
run a script which would start all the copies of the other VMs (that were rsync'd the night before).
Thanks for any guidance.
Brent