Thanks, guys. I forgot the IOPS. So since I have 100disks, the total IOPS=100X100=10K. For the 4+2 erasure, one disk fail, then it needs to read 5 and write 1 objects.Then the whole 100 disks can do 10K/6 ~ 2K rebuilding actions per seconds. While for the 100X6TB disks, suppose the object size is set to 4MB, then 6TB/4MB=1.25 million objects. Not considering the disk throughput IO or CPUs, fully rebuilding takes: 1.25M/2K=600 seconds? Best, Feng On Thu, May 9, 2019 at 10:23 AM Janne Johansson <icepic.dz@xxxxxxxxx> wrote: > > Den tors 9 maj 2019 kl 16:17 skrev Marc Roos <M.Roos@xxxxxxxxxxxxxxxxx>: >> >> >> > Fancy fast WAL/DB/Journals probably help a lot here, since they do >> affect the "iops" >> > you experience from your spin-drive OSDs. >> >> What difference can be expected if you have a 100 iops hdd and you start >> using >> wal/db/journals on ssd? What would this 100 iops increase to >> (estimating)? >> > > I don't know, there is a factor of reading objects which won't get lots of perf from > WAL/DB/Journals at all, only the destination writes, and also the relative sizes of > the WAL/Journals are relevant since they need to be large enough to allow the > drive to flush out data (albeit in a nicer order with larger IOs presumably) or you will > just have nice IOPS for a while and then fall back to spin-drive speeds as the > WAL/Journal gets filled and need to wait for the drives anyhow. > > -- > May the most significant bit of your life be positive. _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |