The use case is for KVM RBD volumes. Our enviroment will be 80% random reads/writes probably 40/60 or 30/70 is a good estimate. All 4k-8k IO sizes. We currently run on a Nimble Hybrid array which runs in the 5k-15k IOPS range with some spikes up to 20-25k IOPS (Capable of 100k iops per Nimble).
Its also worth mentioning that I plan on splitting this cluster between our two sites. We have a dedicated dark fiber connection which is 8x 10GB links between the two sites (roughly .4 or .5 ms latency). I was leaning towards 40G because our existing HP 5900 switch's which we use to light our dark fiber also have 4x 40GB ports which I could use to hang the Ceph cluster network from. This would be a 4/2 setup, keeping 2 copies at each site. I currently have a much lower end ceph cluster using all spinning rust with 3 nodes at each site (48 OSD's). So far that is working out really well, but the data going on that cluster is written than almost never touched again. The above spinning rust cluster is running just simple 10G SFP+. _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-le...@ceph.io
