The 27/06/12, Nathan Stratton wrote: > What is considered half-decent? I have a 8 cluster > distribute+replicate setup and I am getting about 65MB/s and about > 1.5K IOPS. Considering that I am only using a single two disk SAS > strip in each host I think that is not bad. Hum, looking at your at your latter mail I would have expect better performance, too. > Also check out oVirt, it integrates with Gluster and provides HA. I already know the existance of oVirt. I'll take a further look at it. > >>2. We still didn't decide what physical network to choose between FC, FCoE > >>and Infiniband. > > > >Have you ruled out 10G ethernet? If so, why? > > I agree, we went all 10GBase-T. We excluded ethernet due to searches on the web. It appeared that ethernet has bad latency. > >>3. Would it be better to split the Glusterfs namespace into two gluster > >>volumes (one for each hypervisor), each running on a Glusterfs server > >>(for the normal case where all servers are running)? > > > >I don't see how that would help - I expect you would mount both volumes on > >both KVM nodes anyway, to allow you to do live migration. > > Yep In the usual case, correct me if I'm wrong, the traffic from all the Glusterfs client (KVM hosts, here) goes to the same Glusterfs server: KVM 1 <----> Glusterfs server A KVM 2 <----> Glusterfs server A If I split out the cluster in two parts, I would expect to be able to distribute the network traffic like this: KVM 1 <----> Glusterfs server A KVM 2 <----> Glusterfs server B And so, having better performance while still beeing HA-compliant. In this splitted mode I think it would be possible to: - handle hypervisors workload (mostly processor) by live-migrating VM. - handle Glusterfs server workload (mostly disks IO) by exporting a VM from one side and importing it in the other Glusterfs server. -- Nicolas Sebrecht
