[ceph-users] optimize bluestore for random write i/o

Stefan Priebe - Profihost AG s.priebe at profihost.ag
Tue Mar 5 03:26:01 PST 2019


Am 05.03.19 um 10:05 schrieb Paul Emmerich:
> This workload is probably bottlenecked by rocksdb (since the small
> writes are buffered there), so that's probably what needs tuning here.

while reading:
https://www.flashmemorysummit.com/English/Collaterals/Proceedings/2018/20180807_INVT-101A-1_Meredith.pdf

there is also a mention of:
osd_min_pg_log_entries = 10
osd_max_pg_log_entries = 10
osd_pg_log_dups_tracked = 10
osd_pg_log_trim_min = 10

which sounds like the backported feature of ceph 12.2.11 ot limit ph log
entries. Can anybody confirm?

They also changed:
bluestore_rocksdb_options



> Paul
> 


More information about the ceph-users mailing list