From owner-freebsd-current Sat Nov 16 17:34:20 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id RAA03869 for current-outgoing; Sat, 16 Nov 1996 17:34:20 -0800 (PST) Received: from hauki.clinet.fi (root@hauki.clinet.fi [194.100.0.1]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id RAA03864 for ; Sat, 16 Nov 1996 17:34:14 -0800 (PST) Received: from cantina.clinet.fi (root@cantina.clinet.fi [194.100.0.15]) by hauki.clinet.fi (8.7.6/8.6.4) with ESMTP id DAA27687; Sun, 17 Nov 1996 03:33:58 +0200 (EET) Received: (hsu@localhost) by cantina.clinet.fi (8.7.5/8.6.4) id DAA16989; Sun, 17 Nov 1996 03:33:57 +0200 (EET) Date: Sun, 17 Nov 1996 03:33:57 +0200 (EET) Message-Id: <199611170133.DAA16989@cantina.clinet.fi> From: Heikki Suonsivu To: Jaye Mathisen Cc: freebsd-current@freebsd.org In-reply-to: Jaye Mathisen's message of 16 Nov 1996 22:46:33 +0200 Subject: Re: I think our fsck is still funky. Organization: Clinet Ltd, Espoo, Finland References: Sender: owner-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk In article Jaye Mathisen writes: 2.2-current, supped this morning. P120, 2940, Conner 4GB disk. Copy lots of small files to the disk in lots of directories (like news). Begin an rm while copying. Simulate failure by pulling plug. Unsimulate failure by re-inerting power cord. Reboot, fsck disk. Mount, and try again. I get crashes in dup_inode, and some ffs_alloc stuff. newfs the disk and it's fine. Me thinks that fsck is still missing some sanity checks. Since end of August (I have kernel compiled 19th) freebsd has been corrupting disks repeatably with -current. For us symptom is always the same, crash, and fsck fails on number of DUP/BAD inodes (random sequences of 8 corrupted inodes around the disk). It seems that couple of blockfulls of inodes get written over, usually files which were used at the time. On news server we get crash too quickly that any later kernel would be usable, I'm still using the old one. At least two people have had same problem, but it happens on IDE disks (quantum bigfoots) and 2.1.5 (!), the hardware is - cyrix p150+, 2*2.5G bigfoot, 64M, stealth 64, adaptek 2940, 5G hawk, DAT, other machine P133, 32M, 1G IDE, Stealth 64, no SCSI. The first machine is server, the latter one X terminal. Originally the latter machine was running 2.1R reliably (for 7 months) with P133, 128M, 1G IDE, Adaptek 2940, 1G SCSI, 4G hawk, DAT. This does not quite fit the picture, if it is scsi problem, as corrupted disks are IDE. For more info ask ylo@ssh.fi We only have seen corruption on SCSI disks but that may be because we generally use SCSI for anything involving any real work, only routers and X-terminals have IDE disks. -- Heikki Suonsivu, T{ysikuu 10 C 83/02210 Espoo/FINLAND, hsu@clinet.fi mobile +358-40-5519679 work +358-9-43542270 fax -4555276