On Friday 29 October 2010 20:31:03 wang xuchen wrote: > As I said in my first email. dd test on raw SSD device reaches almost > 100MB/s with 4k block size(which is almost 26,000 IOPS). I apologize that I > haven`t mentioned the CPU utilization which is surprisingly low, all worker > threads and receive threads consume less than 1% CPU individually. That why > how I come up with the conclusion either my configuration is wrong or DRBD > replication too slow. > > > Any suggestion is appreciated. > > Ben > > Commit yourself to constant self-improvement > > On Fri, Oct 29, 2010 at 1:01 PM, J. Ryan Earl <[email protected]> wrote: > > On Fri, Oct 29, 2010 at 11:13 AM, Julien Escario <[email protected]>wrote: > >> Someone answered me that these preformances are perfectly normal is dual > >> master configuration. > > > > Someone told you wrong. > > > >> Seems to be due to the network latency (first server I/O + network > >> latency + second server I/O + network latency (ACK)) > > > > You should still be able to get much better than 97MB/sec. I get over > > 800MB/sec with a dual-primary with a backing store capable of ~950 > > MB/sec. > > > >> I finally decided that DRBD is unusable in dual primary setup because > >> the performance drop. > > > > You were using 8.0.16, you should have tried with 8.3.x, 8.3 seems to do > > dual-primary a lot better in general. > > > > -JR > > > > _______________________________________________ > > drbd-user mailing list > > [email protected] > > http://lists.linbit.com/mailman/listinfo/drbd-user
And your DRBD network? _______________________________________________ drbd-user mailing list [email protected] http://lists.linbit.com/mailman/listinfo/drbd-user
