: [ceph-users] Switching from tcmalloc
It would be really interesting if you could give jemalloc a try.
Originally tcmalloc was used to get around some serious memory
fragmentation issues in the OSD. You can read the original bug tracker
entry from 5 years ago here:
http://tracker.ceph.com/issues
IMHO there must be tested in different glibc. Old glibc has optional
experimental threaded extensions for malloc, default disabled (and have no
options to enable even in Gentoo without hack, may be some distros was compiled
so - don't know). But now this malloc features mostly ON by default, so
...@lists.ceph.com] On Behalf Of Jan
Schermer
Sent: Wednesday, June 24, 2015 10:54 AM
To: Ben Hines
Cc: ceph-users@lists.ceph.com
Subject: Re: [ceph-users] Switching from tcmalloc
We did, but I don’t have the numbers. I have lots of graphs, though. We were
mainly trying to solve the CPU usage, since
Our first thought was jemalloc when we became aware of the issue, but that one
requires support in code which is AFAIK not present in Dumpling. Am I right?
We did try simply preloading jemalloc when starting OSD and that experient
ended with SIGSEGV within minutes, we didn’t investigate it any
Can you guess when we did that?
Still on dumpling, btw...
http://www.zviratko.net/link/notcmalloc.png
http://www.zviratko.net/link/notcmalloc.png
Jan___
ceph-users mailing list
ceph-users@lists.ceph.com
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
Did you see what the effect of just restarting the OSDs before using
tcmalloc? I've noticed that there is usually a good drop for us just
by restarting them. I don't think it is usually this drastic.
-
Robert LeBlanc
GPG
...@lists.ceph.com] On Behalf Of Jan
Schermer
Sent: Wednesday, June 24, 2015 10:54 AM
To: Ben Hines
Cc: ceph-users@lists.ceph.com
Subject: Re: [ceph-users] Switching from tcmalloc
We did, but I don’t have the numbers. I have lots of graphs, though. We were
mainly trying to solve the CPU usage, since our
-users-boun...@lists.ceph.com] On Behalf Of Jan
Schermer
Sent: Wednesday, June 24, 2015 10:54 AM
To: Ben Hines
Cc: ceph-users@lists.ceph.com
Subject: Re: [ceph-users] Switching from tcmalloc
We did, but I don’t have the numbers. I have lots of graphs, though. We were
mainly trying to solve the CPU
Did you do before/after Ceph performance benchmarks? I dont care if my
systems are using 80% cpu, if Ceph performance is better than when
it's using 20% cpu.
Can you share any scripts you have to automate these things? (NUMA
pinning, migratepages)
thanks,
-Ben
On Wed, Jun 24, 2015 at 10:25 AM,
We did, but I don’t have the numbers. I have lots of graphs, though. We were
mainly trying to solve the CPU usage, since our nodes are converged QEMU+CEPH
OSDs, so this made a difference. We were also seeing the performance capped on
CPUs when deleting snapshots of backfilling, all this should
We already had the migratepages in place before we disabled tcmalloc. It didn’t
do much.
Disabling tcmalloc made immediate difference but there were still spikes and
the latency wasn’t that great. (CPU usage was)
Migrating memory helped a lot after that - it didn’t help (at least not the
There were essentialy three things we had to do for such a drastic drop
1) recompile CEPH —without-tcmalloc
2) pin the OSDs to a set of a specific NUMA zone - we had this for a long time
and it really helped
3) migrate the OSD memory to the correct CPU with migratepages
- we will use cgroups
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
From what I understand, you probably got most of your reduction from
co-locating your memory to the right NUMA nodes. tcmalloc/jemalloc
should be much higher in performance because of how they hold memory
in thread pools (less locking to allocate
13 matches
Mail list logo