Hi, On Thu, Nov 11, 2021 at 3:26 PM Janne Johansson <icepic.dz@xxxxxxxxx> wrote: > Den tors 11 nov. 2021 kl 13:54 skrev Zakhar Kirpichenko <zakhar@xxxxxxxxx > >: > > I'm still trying to combat really bad read performance from HDD-backed > > replicated pools, which is under 100 MB/s most of the time with 1 thread > > and QD=1. I don't quite understand why the reads are that slow, i.e. much > > (doing a single-thread-single-client test on a cluster ..) > Doing a single-thread single-client test on a cluster is totally fine if single-thread single-client performance is important. > Also, if you map an RBD like when you have the rbd image as an 40G > qemu drive for a VM guest, it will get split into 4M pieces anyhow, so > if the guest decides to read its drive from 0 -> end it will fire off > 10000 read requests, spread out over the cluster and all OSDs on which > the pool is placed, so you get load sharing. > I guess reading the guest disk from 0 to end sequentially doesn't do any load sharing. Reading multiple parts of the guest disk simultaneously has a good chance to. experience. Also, 100MB/s from a spin drive over a network isn't all > that bad, given QD=1, since the double network latency will be there > for linear reads for every block you ask for. If the turn-around times > over the network is 1ms, then 1000 x IO-size is all you could hope for > optimally at QD=1. > That's not how it works in practice, otherwise I'd be getting (1000/RTT)*IO-size from SSD drives as well, but that is not the case. > So while it is easy to visualize a huge improvement by asking more IO > from what you imagine is idle drives, the normal cluster will be > spreading tons and tons of IO all kinds of ways all the time, making > the server IO queue deeper is probably not going to improve the sum of > all IO that goes to clients. > Yes, I realize this. I do have a quiet cluster though :-) Z _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |