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 visibly on graphs) when tcmalloc was used - it’s overhead was so large NUMA didn’t matter at all. But we are running Dumpling, so it is possible other bottlenecks that were resolved in later (Giant) releases would once again overshadow the gain we got from disabling tcmalloc or there would be regression from disabling it. … or our setup/workload is somehow completely different from what somebody else has? Jan > On 24 Jun 2015, at 19:41, Robert LeBlanc <rob...@leblancnet.us> wrote: > > -----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 memory) and they try much harder to reuse dirty > free pages so memory stays within the thread again reducing locking for > memory allocations. > > I would do some more testing along with what Ben Hines mentioned about > overall client performance. > > - ---------------- > Robert LeBlanc > GPG Fingerprint 79A2 9CA4 6CC4 45DD A904 C70E E654 3BB2 FA62 B9F1 > > On Wed, Jun 24, 2015 at 11:25 AM, Jan Schermer wrote: > 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 in the future for this, should make life easier and is > the only correct solution > > It is similiar to the effect of just restarting the OSD, but much better - > since we immediately see hundreds of connections on a freshly restarted OSD > (and in the benchmark the tcmalloc issue manifested with just two clients in > parallel) I’d say we never saw the raw performance with tcmalloc > (undegraded), but it was never this good - consistently low latencies, much > smaller spikes when something happens and much lower CPU usage (about 50% > savings but we’re also backfilling a lot on the background). Workloads are > faster as well - like reweighting OSDs on that same node was much (hundreds > of percent) faster. > > So far the effect has been drastic. I wonder why tcmalloc was even used when > people are having problems with it? The glibc malloc seems to work just fine > for us. > > The only concerning thing is the virtual memory usage - we are over 400GB VSS > with a few OSDs. That doesn’t hurt anything, though. > > Jan > > > On 24 Jun 2015, at 18:46, Robert LeBlanc wrote: > > - -----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 Fingerprint 79A2 9CA4 6CC4 45DD A904 C70E E654 3BB2 FA62 B9F1 > > On Wed, Jun 24, 2015 at 2:08 AM, Jan Schermer wrote: > 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 <mailto:ceph-users@lists.ceph.com> > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com > <http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com> > > > - -----BEGIN PGP SIGNATURE----- > Version: Mailvelope v0.13.1 > Comment: https://www.mailvelope.com <https://www.mailvelope.com/> > > wsFcBAEBCAAQBQJVit75CRDmVDuy+mK58QAAmjcP/jU+wyohdwKDP+FHDAgJ > DcqdB5aPG2AM79iLcYUub5bQjdNJpcWN/hyZcNdF3aSzEV3aY6jIqu9OpOIB > c2fIzfGOoczzW/FEf7qKRVGpxaQL21Sw1LpwMEscNe0ETz9HMHoaAnBO9IFn > nUEOCdEpRBO5W1rWwNAx9EVnOUPklb7vVEpY23sgtHhQSprb9oeO8D99AMRz > /RhdHKlRDgHBjun/stCiR6lFuvBUx0GBmyaMuO5rfsLGRIkySLv++3CLQI6X > NCt/MjYwTTNNfO/y/MjkiV/j+Cm1G1lcjlgbDjilf7bgf8/7W2vJa1sMtaA4 > xJL+PpZxiKcGSdC96B+EBYxLhLcwsNpbfq7uxQOkIspa66mkIMAVzJgt4DFL > Ca+UY3ODA26VtWF5U/hkdupgld+YSxXTyJakeShrBSFAX0a4cygV9Ll7SIhO > IDS+0Mbur0IGzIWRgtCQhRXsc7wn3IoIovqe8Nfk4xupeoK2P5UHO1rW9pWy > Jwj5PXieDqxgx8RKlulN1bCbSgTaEdveTiqqVxlnM9L0MhgesuB8vkpHbsqn > mYJHNzU7ghU89xLnRuia9rBlpjw4OzagfowAJTH3UnaO67kxES+IWO8onQbN > RhY0QR5cB5rVSjYkzzlsuLM17fQPcT8++yMarKdsrr6WIGppXUFFdATAqIaY > DHD1 > =goL4 > - -----END PGP SIGNATURE----- > > > -----BEGIN PGP SIGNATURE----- > Version: Mailvelope v0.13.1 > Comment: https://www.mailvelope.com <https://www.mailvelope.com/> > > wsFcBAEBCAAQBQJViuu0CRDmVDuy+mK58QAASzoQAIf4Lj/jA2yl2XMS7RAW > FmgK8rsf2iyzg6UQMmobFw0oWTb/0T4AscXlZIE7dhGUi6m6UHWBPB7P6YBZ > UQ2eJqzcaK1Jf/flfTZajWB2z2CSYpuwbPYaQ8SqKoyauEjKgD092/LUfKL3 > TP5z7SdhZ8/HmzT2qFUdYuAQ+WvD3rgdJtkblFgItM+bqKmhibiZr3KHzXoU > j5Ob61AsR6/s3hgWJ09uAghqB8SNsxJ0u7R5RnaiS2VWkHSHTrdiTwd/ONlL > anBnKljTgkCSqS3RoPVB74qlqhDxlDnwRYvKrxurcikaI3tZ17xt4UvCc9yP > RRH6M8aU1+7itOxu8DyOeZ+9Ev5/H6i5LwtrnN2pHaN9s0tWRKwzt5HQEYhE > ceoyui+EtpN8zzqs9ryIGvHL3KB1bmL+0WWO4RlT8NwodsSge3Yga8KUMa07 > 8+dh0VGUywGEmxMg2VWPyvKf/keOiWHHi4UDJRgXJdnBjH/+4Yebva7TJ2b9 > Ch0r8JL00nbJCBb78dvw59XiFUJBFT5WfgItmbfjX2SI+srFaDXFKtGSjnFi > MK4gE7DA70tKgP+xwpw3Eou7rDzxogqxnV54BlNzvbokbfiDAZ/ARL7CtC/1 > SnBxzEaliaJnBHKSgwOyP9sxz+QKMxty2ZTSmCnBUxKRK9O2hNSzFf6+1heT > yQ3L > =DreJ > -----END PGP SIGNATURE-----
_______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
