On Sat, Dec 13, 2008 at 6:06 PM, lists@xxxxxxxxxxxx <lists@xxxxxxxxxxxx> wrote: >> By now I'm able to suspend all vm on a node, stop the vmware server >> that hosted them, start another vmware server on another node and >> resume them there. > > This could work also, though it's not redundancy, it's more of a backup solution. > The issue with this is that win machines seem to complain about being moved > from one server to another and require re-activating. > > It's possible that this was because I might have changed a hardware setting in > the guest which caused this to happen, rather than the move from one VM > server to another. I'll have to test that. Maybe, btw I had another issue with linux guests and vmware server 2.0: after moving them to another node, vmware required to register them again. I had to play here with .vmx options and moving some /etc/vmware files to clustered gfs to avoid this. And vmware server 1.0.x seems more stable here. For relocating the guests to another cluster node I did a custom script that first tries to suspend the guests, if it fails, it powers off them, then stops the vmware daemons, and everything is ready for a new node vmware server to start with the same guests. Regards, Diego. -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
