Hi! I am trying to find if Ceph has any QoS settings that apply
per-client. I would like to be able, say, to have different QoS settings
for RBD clients named "nova" and "cinder" and different again from an
RGW client named "enduser".
While researching this, I found amongst other things, the following
references, which do not appear to be what I want:
- https://docs.ceph.com/en/reef/dev/osd_internals/mclock_wpq_cmp_study/
and
https://docs.ceph.com/en/reef/rados/configuration/osd-config-ref/#dmclock-qos
. This appears to be instead about the balancing of user-iops against
ceph-internal iops, not balancing clients against each other.
(aside: somebody should rewrite this paragraph, I find it totally
incomprehensible:
https://docs.ceph.com/en/reef/rados/configuration/osd-config-ref/#subtleties-of-mclock
)
- https://docs.ceph.com/en/latest/rbd/rbd-config-ref/#qos-settings and
https://tracker.ceph.com/issues/36191 This seems to be about setting
limits on the RBD pools as a whole, or on separate images inside them.
However separate clients each still get to use up to the limits.
Is there anything else that I have not found so far but which is about
balancing individual clients across all services?
--
Olaf Seibert
Site Reliability Engineer
SysEleven GmbH
Boxhagener Straße 80
10245 Berlin
T +49 30 233 2012 0
F +49 30 616 7555 0
https://www.syseleven.de
https://www.linkedin.com/company/syseleven-gmbh/
Current system status always at:
https://www.syseleven-status.net/
Company headquarters: Berlin
Registered court: AG Berlin Charlottenburg, HRB 108571 Berlin
Managing directors: Andreas Hermann, Jens Ihlenfeld, Norbert Müller,
Jens Plogsties
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
