From owner-freebsd-current@FreeBSD.ORG Sun Oct 15 08:35:39 2006 Return-Path: X-Original-To: current@freebsd.org Delivered-To: freebsd-current@FreeBSD.ORG Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 0070616A416 for ; Sun, 15 Oct 2006 08:35:39 +0000 (UTC) (envelope-from doublef-ctm@yandex.ru) Received: from smtp1.yandex.ru (smtp1.yandex.ru [213.180.223.87]) by mx1.FreeBSD.org (Postfix) with ESMTP id D521F43D6B for ; Sun, 15 Oct 2006 08:35:20 +0000 (GMT) (envelope-from doublef-ctm@yandex.ru) Received: from [83.239.189.66] ([83.239.189.66]:28111 "EHLO shark" smtp-auth: "doublef-ctm" TLS-CIPHER: TLS-PEER-CN1: ) by mail.yandex.ru with ESMTP id S2078156AbWJOIfN (ORCPT ); Sun, 15 Oct 2006 12:35:13 +0400 X-Comment: RFC 2476 MSA function at smtp1.yandex.ru logged sender identity as: doublef-ctm Received: by shark (Postfix, from userid 1000) id 310D9177AD; Sun, 15 Oct 2006 12:35:11 +0400 (MSD) Date: Sun, 15 Oct 2006 12:35:11 +0400 From: Sergey Zaharchenko To: current@freebsd.org Message-ID: <20061015083511.GA13203@shark.localdomain> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="WIyZ46R2i8wDzkSu" Content-Disposition: inline X-Listening-To: Silence User-Agent: Mutt/1.5.11 Cc: Subject: Processes in "ufs" state (PR kern/104406) X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sun, 15 Oct 2006 08:35:39 -0000 --WIyZ46R2i8wDzkSu Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hello list! I've stumbled across this (probable) bug when doing some number-crunching. Processes which try to access the filesystem hang at some point when another process has been gobbling up CPU time for a sufficiently long time. Interestingly, processes which have already opened some files before the `hang' can apparently read/write them without hindrance (iostat shows some disk activity). Interactive processes run well too. When the number-cruncher is `kill -STOP'ped and `kill -CONT'ed, everything runs well for another while. It's reliably and frequently reproducible here both on recent and late August CURRENT, the kernel is GENERIC, the hardware is basic (no RAID, etc.). Could anyone experienced enough try to look into this and/or give advice for debugging the problem? I have ddb available, but I'm not sure about where to look. http://www.freebsd.org/cgi/query-pr.cgi?pr=3D104406 Thanks in advance, --=20 DoubleF No virus detected in this message. Ehrm, wait a minute... /kernel: pid 56921 (antivirus), uid 32000: exited on signal 9 Oh yes, no virus:) --WIyZ46R2i8wDzkSu Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (FreeBSD) iD8DBQFFMfK/wo7hT/9lVdwRAhDVAJ0X0G+oUMkLS16Rt/SQCD4f/ngMMACdG5WC uC+kCnc8ADQrb7kkcy5D5Uk= =uLGY -----END PGP SIGNATURE----- --WIyZ46R2i8wDzkSu--