Received: from malur.postgresql.org ([217.196.149.56]) by arkaria.postgresql.org with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1jfv1g-0006Hk-PT for pgsql-sql@arkaria.postgresql.org; Tue, 02 Jun 2020 00:43:32 +0000 Received: from localhost ([127.0.0.1] helo=malur.postgresql.org) by malur.postgresql.org with esmtp (Exim 4.92) (envelope-from ) id 1jfv1f-0000PS-L5 for pgsql-sql@arkaria.postgresql.org; Tue, 02 Jun 2020 00:43:31 +0000 Received: from magus.postgresql.org ([2a02:c0:301:0:ffff::29]) by malur.postgresql.org with esmtps (TLS1.3:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1jfv1f-0000PL-EH for pgsql-sql@lists.postgresql.org; Tue, 02 Jun 2020 00:43:31 +0000 Received: from li1929-156.members.linode.com ([172.104.219.156] helo=ln-1.useunix.net) by magus.postgresql.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1jfv1c-0005hY-15 for pgsql-sql@postgresql.org; Tue, 02 Jun 2020 00:43:30 +0000 Received: from wcuddy by ln-1.useunix.net with local (Exim 4.92.3) (envelope-from ) id 1jfv1Z-0003Eg-On for pgsql-sql@postgresql.org; Tue, 02 Jun 2020 00:43:25 +0000 Date: Tue, 2 Jun 2020 00:43:25 +0000 From: Wayne Cuddy To: PostgreSQL Subject: autovacuum processes not spawning Message-ID: <20200602004325.GL20190@ln-1.useunix.net> Mail-Followup-To: PostgreSQL MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.10.1 (2018-07-13) List-Id: List-Help: List-Subscribe: List-Post: List-Owner: List-Archive: Precedence: bulk Version: PGSQL-TimescaleDB 9.6 OS: Debian Linux, Kernel 4.9.0 Been running happily on a fairly busy server for quite some time. Aside from the occasional hardware issues it's been very stable. Last Friday we noticed we had performance issues with tables that are usually never a problem. Nothing changed WRT our application with is quite predictable. We determined that on Thursday night autovacuum processes were no longer starting. Our server is configured to log whenever AV runs start as well as log all durations. The autovacuum launcher process was still running and there were no errors in the log. We could manually initially vacuums and they would run to completion without error. Today we bounced the database process and it's now running autovacuums at regular intervals and performance is what we expect. Has anyone else had autovacuums just stop launching? Is there a way to restart the autovacuum launcher without restarting the entire database? I'm not positive it's the launcher but I have nothing else to blame at this point. Wayne