Hi Nick, I found the graph very useful explaining the concept. thx for sharing. I'm currently planning to setup a new cluster and wanted to get low latency by using, 2U server, 6xIntel P3700 400GB for journal and 18x1.8TB Hitachi Spinning 10k SAS. My OSD:Journal ratio would be 3:1. All over 10Gbit copper network. For CPU I'm thinking of E5-2667v4 (Single socket, 8 cores, 3.2Ghz) Replica size=3, and 5 hosts. Do you think it is possible to get a low direct sync write latency 3-5ms-avg with a configuration like this, or would the spinning drives actually slow down this whole design? thx will On Sat, Nov 12, 2016 at 3:33 AM, Nick Fisk <nick@xxxxxxxxxx> wrote: > Hi All, > > I've recently put together some articles around some of the performance testing I have been doing. > > The first explores the high level theory behind latency in a Ceph infrastructure and what we have managed to achieve. > > http://www.sys-pro.co.uk/ceph-write-latency/ > > The second explores some of results we got from trying to work out how much CPU a Ceph IO uses. > > http://www.sys-pro.co.uk/how-many-mhz-does-a-ceph-io-need/ > > I hope they are of interest to someone. > > I'm currently working on a couple more explaining the choices behind the hardware that got us 700us write latency and what we > finally built. > > Nick > > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |