| From: | Robert Haas <robertmhaas(at)gmail(dot)com> |
|---|---|
| To: | Alexander Kuzmenkov <a(dot)kuzmenkov(at)postgrespro(dot)ru> |
| Cc: | Amit Kapila <amit(dot)kapila16(at)gmail(dot)com>, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>, Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, Andres Freund <andres(at)anarazel(dot)de>, Alexander Korotkov <a(dot)korotkov(at)postgrespro(dot)ru>, Simon Riggs <simon(at)2ndquadrant(dot)com>, Dilip Kumar <dilipbalaut(at)gmail(dot)com>, Jeff Davis <pgsql(at)j-davis(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>, Greg Stark <stark(at)mit(dot)edu>, Andres Freund <andres(at)2ndquadrant(dot)com>, Rajeev rastogi <rajeev(dot)rastogi(at)huawei(dot)com>, Markus Wanner <markus(at)bluegap(dot)ch>, Ants Aasma <ants(at)cybertec(dot)at>, Bruce Momjian <bruce(at)momjian(dot)us>, obartunov <obartunov(at)postgrespro(dot)ru>, Teodor Sigaev <teodor(at)postgrespro(dot)ru>, Borodin Vladimir <root(at)simply(dot)name> |
| Subject: | Re: Proposal for CSN based snapshots |
| Date: | 2017-10-05 21:34:39 |
| Message-ID: | CA+TgmoY6LTMhsQurr2xjYrK+-Ov8jF07OsfvojFRYGHacga2XQ@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Mon, Sep 25, 2017 at 10:17 AM, Alexander Kuzmenkov
<a(dot)kuzmenkov(at)postgrespro(dot)ru> wrote:
> Here is some news about the CSN patch.
>
> * I merged it with master (58bd60995f), which now has the clog group update.
> With this optimization, CSN is now faster than the master by about 15% on
> 100 to 400 clients (72 cores, pgbench tpcb-like, scale 500). It does not
> degrade faster than master as it did before. The numbers of clients greater
> than 400 were not measured.
Hmm, that's gratifying.
> * Querying for CSN of subtransactions was not implemented in the previous
> version of the patch, so I added it. I tested the performance on the
> tpcb-like pgbench script with some savepoints added, and it was
> significantly worse than on the master. The main culprit seems to be the
> ProcArrayLock taken in GetSnapshotData, GetRecentGlobalXmin,
> ProcArrayEndTransaction. Although it is only taken in shared mode, just
> reading the current lock mode and writing the same value back takes about
> 10% CPU. Maybe we could do away with some of these locks, but there is some
> interplay with imported snapshots and replication slots which I don't
> understand well. I plan to investigate this next.
That's not so good, though.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Joshua D. Drake | 2017-10-05 21:37:21 | search path security issue? |
| Previous Message | Sean Chittenden | 2017-10-05 21:34:37 | Re: Re: [PATCH] BUG #13416: Postgres =?utf-8?Q?>=3D_?=9.3 doesn't use optimized shared memory on Solaris anymore |