Hello, I am wondering if there is anyone on list that is successfully serving virtual machine images (Xen, KVM, VMWare, etc.) over GlusterFS. I am concerned about difficulties the interaction with GlusterFS and large files that are held open for large periods of time (except when the VM is rebooted). My main concern is with replication and failover. Is it currently possible to setup relatively transparent failover of a brick or will the VMs all need to be rebooted to "reopen" their disk images when a brick goes down. The setup we were thinking of going with here is client striping + server side AFR. I don't know whether HA will need to be involved in order to make this happen. Also, for the current self-healing mechanism to work my understanding is that the image (file) will need to be closed and reopened, i.e. the VM will need to be rebooted but this may be fixed with the new healing tools I think are coming in 2.0? Thanks, Alexander
