X-Original-To: pgsql-admin-postgresql.org@localhost.postgresql.org Received: from localhost (neptune.hub.org [200.46.204.2]) by svr1.postgresql.org (Postfix) with ESMTP id CB8E3D1D767 for ; Wed, 14 Jan 2004 04:36:54 +0000 (GMT) Received: from svr1.postgresql.org ([200.46.204.71]) by localhost (neptune.hub.org [200.46.204.2]) (amavisd-new, port 10024) with ESMTP id 11310-10 for ; Wed, 14 Jan 2004 00:36:23 -0400 (AST) Received: from sss.pgh.pa.us (unknown [192.204.191.242]) by svr1.postgresql.org (Postfix) with ESMTP id BE8E6D1D729 for ; Wed, 14 Jan 2004 00:36:22 -0400 (AST) Received: from sss2.sss.pgh.pa.us (tgl@localhost [127.0.0.1]) by sss.pgh.pa.us (8.12.10/8.12.10) with ESMTP id i0E4aK19029722; Tue, 13 Jan 2004 23:36:21 -0500 (EST) To: Michael Adler Cc: pgsql-admin@postgresql.org Subject: Re: vacuum slowed by syslogd In-reply-to: <20040114031030.GA18188@pobox.com> References: <20040114031030.GA18188@pobox.com> Comments: In-reply-to Michael Adler message dated "Tue, 13 Jan 2004 22:10:31 -0500" Date: Tue, 13 Jan 2004 23:36:19 -0500 Message-ID: <29721.1074054979@sss.pgh.pa.us> From: Tom Lane X-Virus-Scanned: by amavisd-new at postgresql.org X-Archive-Number: 200401/113 X-Sequence-Number: 11976 Michael Adler writes: > On many occasions, I've noticed that some PostgreSQL activity takes > far longer than it previously did and that disabling syslogd addresses > the symptoms. > ... > This evidence normally indicates a name resolution issue, but I'm not > sure how to test for that beyond using "hostname -v -i". I looked in Red Hat's bugzilla for similar issues, and found a mention that syslogd tries to reverse-lookup the address it gets each message from. A delay there could act as you describe. At least in recent RH releases, you can start syslogd with the "-x" switch to disable this lookup --- does that exist in Debian, and if so does it help? I would expect any comparable resolution delays on the client side to occur only once when a backend is started, not once per message, since we only openlog() once per backend. So a delay on the syslogd side seems a likely theory. Something else you could do to investigate is to test whether syslogd can log messages from other clients during these 10-second delays. If it's hung trying to resolve a name, presumably messages from other clients would get delayed too. regards, tom lane