Hi Iban,
Most of these options fall under the osd section. You can get
descriptions of what they do here:
https://docs.ceph.com/en/latest/rados/configuration/osd-config-ref/
The journal settings are for the old filestore backend and aren't
relevant unless you are using it. Still, you can see those settings here:
https://docs.ceph.com/en/latest/rados/configuration/journal-ref/
Generally we try to pick reasonable middle-of-the-road defaults.
Sometimes they work well and sometimes additional tuning can help. We
do recommend using care though. People often change multiple options at
once, see a performance improvement, and then word spreads and other
people change the same set of things. In fact it might have only been
one of the options that improved things, and the other options that were
changed cause problems that weren't immediately obvious.
Mark
On 4/12/22 06:17, Iban Cabrillo wrote:
Hi,
Following with the performance (mimic, 144 SATA disk 10Gbps network),the [OSD] entry has the default conf there is no tunnig yet.
I see a lot of parameters that can be set :
[OSD]
osd journal size =
osd max write size =
osd client message size cap =
osd deep scrub stride =
osd op threads =
osd disk threads =
osd map cache size =
osd map cache bl size =
osd recovery op priority =
osd recovery max active =
osd max backfills =
osd min pg log entries =
osd max pg log entries =
osd mon heartbeat interval =
ms dispatch throttle bytes =
objecter inflight ops =
osd op log threshold =
osd crush chooseleaf type =
journal max write bytes =
journal max write entries =
journal queue max ops =
journal queue max bytes =
[Client]
rbd cache =
rbd cache size =
rbd cache max dirty =
Could someone provide any advise?
Regards, I
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
