Novel Failover Idea?

[wnpaul]

I hope this is a novel idea and that it is feasible:

If I have two nodes (node1 and node2) with a DRBD shared storage between them, could I copy /etc/qemu-server on each server to /shared_storage/etc/qemu-server-nodeX and then replace the original directory on each node with a symlink to the directory on the shared storage?

This way, in case of failure I could just copy the *.conf files for the VMs on the failed node into the qemu-server directory of the active node, and restart my VMs.

Is there work underway on HA functionality with automatic failover procedures?

Thanks,

Wolf

 

[dietmar]

This way, in case of failure I could just copy the *.conf files for the VMs on the failed node into the qemu-server directory of the active node, and restart my VMs.

never tried, but I assume that works.

Is there work underway on HA functionality with automatic failover procedures?

Yes, we currently work on that.