I have gluster replica setup with 2 nodes, one disk per node. Sustained writes are good enough - 80MB/s, reads are 140/MB/s
But once I start 4+ windows VM's on it I have problems with iowait sitting at 10%, peaking at 20%, this produces noticable pauses in the VM's (VDI usage). Its interesting because the nas, which only had a transfer rate of 30MB/s didn't have this problem, I presume because it was running in RAID10. If I add a SSD cache using bcache, would this help reduce iowait? I'm using pve kernel 3.10-5, which I presume has bcache integrated, as in the mainline kernel. thanks, -- Lindsay _______________________________________________ pve-user mailing list [email protected] http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user
