[ceph-users] luminous vs jewel rbd performance
rafael.lopez at monash.edu
Wed Nov 15 16:59:14 PST 2017
Sorry for the late reply... I have been away on vacation/openstack summit
etc for over a month now and looking at this again.
Yeah the snippet was a bit misleading. The fio file contains small block
jobs as well as big block jobs:
The performance hit is more noticeable on bigblock, I think up to 10x
slower on some runs but as a percentage it seems to affect a small block
workload too. I understand that runs will vary... I wish I had more runs
from before upgrading to luminous but I only have that single set of
results. Regardless, I cannot come close to that single set of results
since upgrading to luminous.
I understand the caching stuff you mentioned, however we have not changed
any of that config since the upgrade and the fio job is exactly the same.
So if I do many runs on luminous throughout the course of a day, including
when we think the cluster is least busy, we should be able to come pretty
close to the jewel result on at least one of the runs or is my thinking
Sage mentioned at openstack that there was a perf regression with librbd
which will be fixed in 12.2.2.... are you aware of this? If so can you send
me the link to the bug?
On 22 September 2017 at 00:31, Mark Nelson <mnelson at redhat.com> wrote:
> Hi Rafael,
> In the original email you mentioned 4M block size, seq read, but here it
> looks like you are doing 4k writes? Can you clarify? If you are doing 4k
> direct sequential writes with iodepth=1 and are also using librbd cache,
> please make sure that librbd is set to writeback mode in both cases. RBD
> by default will not kick into WB mode until it sees a flush request, and
> the librbd engine in fio doesn't issue one before a test is started. It
> can be pretty easy to end up in a situation where writeback cache is active
> on some tests but not others if you aren't careful. IE If one of your
> tests was done after a flush and the other was not, you'd likely see a
> dramatic difference in performance during this test.
> You can avoid this by telling librbd to always use WB mode (at least when
> rbd cache writethrough until flush = false
> On 09/20/2017 01:51 AM, Rafael Lopez wrote:
>> Hi Alexandre,
>> Yeah we are using filestore for the moment with luminous. With regards
>> to client, I tried both jewel and luminous librbd versions against the
>> luminous cluster - similar results.
>> I am running fio on a physical machine with fio rbd engine. This is a
>> snippet of the fio config for the runs (the complete jobfile adds
>> variations of read/write/block size/iodepth).
>> On 20 September 2017 at 16:43, Alexandre DERUMIER <aderumier at odiso.com
>> <mailto:aderumier at odiso.com>> wrote:
>> so, you use also filestore on luminous ?
>> do you have also upgraded librbd on client ? (are you benching
>> inside a qemu machine ? or directly with fio-rbd ?)
>> (I'm going to do a lot of benchmarks in coming week, I'll post
>> results on mailing soon.)
>> ----- Mail original -----
>> De: "Rafael Lopez" <rafael.lopez at monash.edu
>> <mailto:rafael.lopez at monash.edu>>
>> À: "ceph-users" <ceph-users at lists.ceph.com
>> <mailto:ceph-users at lists.ceph.com>>
>> Envoyé: Mercredi 20 Septembre 2017 08:17:23
>> Objet: [ceph-users] luminous vs jewel rbd performance
>> hey guys.
>> wondering if anyone else has done some solid benchmarking of jewel
>> vs luminous, in particular on the same cluster that has been
>> upgraded (same cluster, client and config).
>> we have recently upgraded a cluster from 10.2.9 to 12.2.0, and
>> unfortunately i only captured results from a single fio (librbd) run
>> with a few jobs in it before upgrading. i have run the same fio
>> jobfile many times at different times of the day since upgrading,
>> and been unable to produce a close match to the pre-upgrade (jewel)
>> run from the same client. one particular job is significantly slower
>> (4M block size, iodepth=1, seq read), up to 10x in one run.
>> i realise i havent supplied much detail and it could be dozens of
>> things, but i just wanted to see if anyone else had done more
>> quantitative benchmarking or had similar experiences. keep in mind
>> all we changed was daemons were restarted to use luminous code,
>> everything else exactly the same. granted it is possible that
>> some/all osds had some runtime config injected that differs from
>> now, but i'm fairly confident this is not the case as they were
>> recently restarted (on jewel code) after OS upgrades.
>> ceph-users mailing list
>> ceph-users at lists.ceph.com <mailto:ceph-users at lists.ceph.com>
>> *Rafael Lopez*
>> Research Devops Engineer
>> Monash University eResearch Centre
>> T: +61 3 9905 9118 <tel:%2B61%203%209905%209118>
>> M: +61 (0)427682670 <tel:%2B61%204%2027682%20670>
>> E: rafael.lopez at monash.edu <mailto:rafael.lopez at monash.edu>
>> ceph-users mailing list
>> ceph-users at lists.ceph.com
> ceph-users mailing list
> ceph-users at lists.ceph.com
Research Devops Engineer
Monash University eResearch Centre
T: +61 3 9905 9118
M: +61 (0)427682670 <%2B61%204%2027682%20670>
E: rafael.lopez at monash.edu
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the ceph-users