From owner-freebsd-questions Tue Feb 6 07:49:05 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id HAA19420 for questions-outgoing; Tue, 6 Feb 1996 07:49:05 -0800 (PST) Received: from Sysiphos (Sysiphos.MI.Uni-Koeln.DE [134.95.212.10]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id HAA19397 for ; Tue, 6 Feb 1996 07:48:49 -0800 (PST) Received: by Sysiphos id AA16530 (5.67b/IDA-1.5 for freebsd-questions@freefall.freebsd.org); Tue, 6 Feb 1996 16:47:47 +0100 Message-Id: <199602061547.AA16530@Sysiphos> From: se@zpr.uni-koeln.de (Stefan Esser) Date: Tue, 6 Feb 1996 16:47:46 +0100 In-Reply-To: Guy Helmer "Re: News server panics (revisited)" (Feb 6, 9:03) X-Mailer: Mail User's Shell (7.2.6 alpha(2) 7/9/95) To: Guy Helmer Subject: Re: News server panics (revisited) Cc: freebsd-questions@freefall.freebsd.org Sender: owner-questions@FreeBSD.ORG Precedence: bulk On Feb 6, 9:03, Guy Helmer wrote: } Subject: Re: News server panics (revisited) } On Tue, 30 Jan 1996, Guy Helmer wrote: } } > My news server has panic'ed twice now in the past two weeks with } > "ffs_alloccg: map corrupted" errors. Does this panic indicate anything } > in particular? I've double-fsck'ed all of the partitions after the last } > crash, and the second fsck was clean on each partition. } } Taking another tack at the question: could these crashes be the result of } having two PCI SCSI controllers in a fairly old (Oct 1994) 486DX2-66 PCI } GW2K? Perhaps the PCI SCSI controllers are hogging the PCI bus and not } giving the PCI IDE disk controller enough bandwidth? The system is } crashing only during expires. INN was built with ACT_STYLE READ. Is this a Saturn based motherboard ? If yes: There might have still been first generation Saturn chip sets been used, which had a few bugs if used with PCI write buffers and burst mode. The Saturn II (== rev. 4) i sknown to work reliably ... I've had other reports of stability problems before, and the system worked flawless after switching off PCI burst mode in the PCI BIOS setup. The failures occur only under high load, as seems to be the case for you, too. Send VERBOSE boot messages if you are not siure about the chip set in your system, and I'll try to identify it and let you know whether it is known buggy ... Regards, STefan -- Stefan Esser, Zentrum fuer Paralleles Rechnen Tel: +49 221 4706021 Universitaet zu Koeln, Weyertal 80, 50931 Koeln FAX: +49 221 4705160 ============================================================================== http://www.zpr.uni-koeln.de/~se