Hi - I can start using the 64 PGs...as Iam having 10 nodes - with 18 OSDs per node..
On Tue, Jul 16, 2019 at 9:01 PM Janne Johansson <icepic...@gmail.com> wrote: > Den tis 16 juli 2019 kl 16:16 skrev M Ranga Swami Reddy < > swamire...@gmail.com>: > >> Hello - I have created 10 nodes ceph cluster with 14.x version. Can you >> please confirm below: >> Q1 - Can I create 100+ pool (or more) on the cluster? (the reason is - >> creating a pool per project). Any limitation on pool creation? >> >> Q2 - In the above pool - I use 128 PG-NUM - to start with and enable >> autoscale for PG_NUM, so that based on the data in the pool, PG_NUM will >> increase by ceph itself. >> >> > 12800 PGs in total might be a bit much, depending on how many OSDs you > have in total for these pools. OSDs aim for something like ~100 PGs per OSD > at most, so for 12800 PGs in total, times 3 for replication=3 makes it > necessary to have quite many OSDs per host. I guess the autoscaler might be > working downwards for your pools instead of upwards. There is nothing wrong > with starting with PG_NUM 8 or so, and have autoscaler increase the pools > that actually do get a lot of data. > > 100 pools * repl = 3 * pg_num 8 => 2400 PGs, which is fine for 24 OSDs but > would need more OSDs as some of those pools grow in data/objects. > > 100 * 3 * 128 => 38400 PGs, which requires 384 OSDs, or close to 40 OSDs > per host in your setup. That might become a limiting factor in itself, > sticking so many OSDs in a single box. > > -- > May the most significant bit of your life be positive. >
_______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
