Received: from malur.postgresql.org ([217.196.149.56]) by arkaria.postgresql.org with esmtp (Exim 4.84_2) (envelope-from ) id 1eHDYo-0001aT-3m for pgsql-performance@arkaria.postgresql.org; Tue, 21 Nov 2017 18:46:18 +0000 Received: from localhost ([127.0.0.1] helo=malur.postgresql.org) by malur.postgresql.org with esmtp (Exim 4.84_2) (envelope-from ) id 1eHDYn-0005uV-Mv for pgsql-performance@arkaria.postgresql.org; Tue, 21 Nov 2017 18:46:17 +0000 Received: from magus.postgresql.org ([2a02:c0:301:0:ffff::29]) by malur.postgresql.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_CBC_SHA384:256) (Exim 4.84_2) (envelope-from ) id 1eHDYn-0005uL-HN for pgsql-performance@lists.postgresql.org; Tue, 21 Nov 2017 18:46:17 +0000 Received: from sss.pgh.pa.us ([66.207.139.130]) by magus.postgresql.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_CBC_SHA384:256) (Exim 4.89) (envelope-from ) id 1eHDYk-0004oZ-GJ for pgsql-performance@lists.postgresql.org; Tue, 21 Nov 2017 18:46:17 +0000 Received: from sss1.sss.pgh.pa.us (localhost [127.0.0.1]) by sss.pgh.pa.us (8.14.4/8.14.4) with ESMTP id vALIk99b014232; Tue, 21 Nov 2017 13:46:09 -0500 From: Tom Lane To: Robert Haas cc: "Henrik Cednert (Filmlance)" , "pgsql-performance@lists.postgresql.org" Subject: Re: pg_dump 3 times as slow after 8.4 -> 9.5 upgrade In-reply-to: References: <8F33D0C7-E12E-4996-990C-3CF0C5ED0437@filmlance.se> <7442.1511283687@sss.pgh.pa.us> Comments: In-reply-to Robert Haas message dated "Tue, 21 Nov 2017 13:39:26 -0500" MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <14230.1511289969.1@sss.pgh.pa.us> Date: Tue, 21 Nov 2017 13:46:09 -0500 Message-ID: <14231.1511289969@sss.pgh.pa.us> List-Id: List-Help: List-Subscribe: List-Post: List-Owner: List-Archive: Precedence: bulk Robert Haas writes: > On Tue, Nov 21, 2017 at 12:01 PM, Tom Lane wrote: >> Can you get a profile of where the machine is spending its time during the >> dump run? On Linux I'd recommend "perf", but on macOS, hmm ... >> You could use Activity Monitor, but as far as I can see that just captures >> short-duration snapshots, which might not be representative of a 10-hour >> run. XCode's Instruments feature would probably be better about giving >> a full picture, but it has a steep learning curve. > macOS's "sample" is pretty easy to use and produces text format output > that is easy to email. Ah, good idea. But note that only traces one process, so you'd need to first determine whether it's pg_dump or the backend that's eating most of the CPU. Or sample both of them. regards, tom lane