On 02/09/14 19:38, Alexandre DERUMIER wrote: > Hi Sebastien, > >>> I got 6340 IOPS on a single OSD SSD. (journal and data on the same partition). > > Shouldn't it better to have 2 partitions, 1 for journal and 1 for datas ? > > (I'm thinking about filesystem write syncs) > Oddly enough, it does not seem to make much difference. I'm seeing approx 8000 IOPS [1] with configs: - journal and data on same partition - journal on one partition, data on another - journal on one partition of one disk (file not device), data on another partition on another disk I think we are just running into the next software bottleneck - our devices are fast enough to *not* be the limiting factor! Cheers Mark [1] I'm seeing a higher IOPS number from Sebastien's, possibly as my ssd are faster for short lived benchmarks, or for some other hardware based reason (single socket i7 + memory timings a bit better for simple workloads etc), but the basic finding is the same I think.