On 12/22/21 4:23 AM, Marc wrote:
I guess what caused the issue was high latencies on our “big” SSD’s (7TB
drives), which got really high after the upgrade to Octopus. We split them
into 4OSD’s some days ago and since then the high commit latencies on the
OSD’s and on bluestore are gone
Hmm, but this is sort of a work around. Unless there is a known difference between how nautilus osd's write to disks compared to octopus (threads, parallel writes ...??)
Is it not just related to, that you created a new osd (regardless of the size)?
This biggest difference would probably be column family sharding. We've
seen some reports that on certain hardware it can negatively affect
write performance (though we've also seen evidence it dramatically
improves compaction behavior). If adding more OSDs helps, I wonder if
having more aggregate WAL and buffer space is what helps, or just more
parallelization on the write path, or perhaps just more cache overall
(assuming you haven't tweaked the memory target to keep the overall
memory usage the same). It could be a combination of all of these too.
Mark
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
