From owner-freebsd-hackers Sun Jun 18 18:41:27 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id SAA15175 for hackers-outgoing; Sun, 18 Jun 1995 18:41:27 -0700 Received: from rwwa.com (rwwa.com [198.115.177.3]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id SAA15169 for ; Sun, 18 Jun 1995 18:41:21 -0700 Received: from localhost (localhost [127.0.0.1]) by rwwa.com (8.6.9/8.6.9) with SMTP id VAA02970 for ; Sun, 18 Jun 1995 21:44:05 -0400 Message-Id: <199506190144.VAA02970@rwwa.com> X-Authentication-Warning: rwwa.com: Host localhost didn't use HELO protocol X-Mailer: exmh version 1.5.3 12/28/94 To: hackers@freebsd.org Subject: Update: 2.0.5 scrambled filesystems. Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Sun, 18 Jun 1995 21:44:05 -0400 From: Robert Withrow Sender: hackers-owner@freebsd.org Precedence: bulk Problem: scrambled filesystems result from doing: ``find / -type f -print | xargs cat >/dev/null'' I've tried several things (listed shortly) and keep getting the same basic problem (either panics due to scrambled filesystems, or the command completes, but a following ``fsck -n'' shows all kinds of corruption. The curruption is still there after a re-boot. Often the system isn't able to sync out the final disk blocks (it says ``giving up''. The fsck errors seen are always on either the /dos filesystem, the /usr filesystem, or both, but *never* on the / (root) filesystem. Also, booting on dos and running scandisk never shows any problems (and running some of the dos software whose file are mentioned on the fsck output is ok too). What I've tried: (each time, I install the ``minimal'' distribution, reboot, fsck -n, do the above command, and if it completes, fsck -n again. Nothing has seemed to make any difference. Note that the system installs correctly (and fsck -n reports *no* errors) and seems to work fine (like I can install emacs, look at files, etc). Things only seem to turn to poop when I do the above. Since nothing I've tried seems to have any effect, I think I am barking up the wrong URL. Could this be geometry related or something? 1) Kept 24MB and turned off the cache (int and ext). 2) Removed 16MB (leaving 8mb), cache on. 3) "" , cache off. 4) 24 MB and the system bios setup in ``slow'' mode. [was it ever]. 5) 24 MB and the system bios setup in ``fast'' mode. 6) 24 MB and changed out the drive cables (what the heck, it was worth a shot, given the recent discussion stating that 98% of failures are connector related...) More complete hardware description: Motherboard vendor unknown (says P54CS-PIO) with AMI bios, 1.20 3/21/95, 75Mhz Pentium, 24MB 2-1x32 and 2-2x32, 256K cache. On the PCI is a diamond 1Mb S3 based vga card. On the isa is a SB16 clone and a WE8013 clone. The drive is a WDC-AC31200F (2484,16,63) translated to (621,64,63). Disk layout is: wd0s1 /dos 159M wd0s2a / 30M wd0s2b swap 96M wd0s2e /usr 937M Aside from the normal boring isa probe info (which I didn't write down), the pci probe says: pci0:0: vendor=0x1039, device=0x406, class=bridge [not supported] pci0:1: vendor=0x1039, device=0x8, class=bridge [not supported] pci0:7: vendor=0x1095, device=0x640, class=storage [not supported] vga0 rev 0 int a irq11 on pci0:15 pci0: uses 8388608 bytes ff000000 upto ff7fffff Ideas? ----------------------------------------------------------------------------- Robert Withrow, Tel: +1 617 598 4480, Fax: +1 617 598 4430 Net: witr@rwwa.COM R.W. Withrow Associates, 319 Lynnway Suite 201, Lynn MA 01901 USA