From owner-freebsd-hackers Tue Jul 4 18:29:21 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id SAA25993 for hackers-outgoing; Tue, 4 Jul 1995 18:29:21 -0700 Received: from cs.weber.edu (cs.weber.edu [137.190.16.16]) by freefall.cdrom.com (8.6.10/8.6.6) with SMTP id SAA25987 for ; Tue, 4 Jul 1995 18:29:19 -0700 Received: by cs.weber.edu (4.1/SMI-4.1.1) id AA03337; Tue, 4 Jul 95 19:22:20 MDT From: terry@cs.weber.edu (Terry Lambert) Message-Id: <9507050122.AA03337@cs.weber.edu> Subject: Stabikity/Usability of 2.0.5R To: davidg@Root.COM Date: Tue, 4 Jul 95 19:22:18 MDT Cc: freebsd-hackers@freebsd.org In-Reply-To: <199507040248.TAA06097@corbin.Root.COM> from "David Greenman" at Jul 3, 95 07:48:37 pm X-Mailer: ELM [version 2.4dev PL52] Sender: hackers-owner@freebsd.org Precedence: bulk Well, I'm on vacation, so I was playing around with 2.0.5R and some old machines that happen to be where I am vacating at. It seems that 2.0.5R is indeed stable as long as you aren't running a split I/D caching. Then the decompression of the kernel fails. On another note, several machines haven't been tested for stability because 2.0.5R simply will not install (and neither will subsequent snapshots). The hardware in question is running WD 1007 ESDI controllers. The "perfect media" jumper is off. The Install wants to use a translated geometry (the real geometry is 1224/15/35) of something like xxx/15/51. The amusing thing is that when it scans for bad blocks (the drive has to have bad block replacement enables in BSD), it gets an error on every 35th scan. I would suspect something wierd with the disk, but 1.1.5.1 sails right on. So it's clearly in the disklabel stuff. Interestingly, the "every 35th try" error on the scan seems to indicate (to me, at least) that the drive is not being accessed linearly; apparently the adjacency of sectors is being miscalculated and it's skipping all over the disk. There was a complaint about large amounts of drive noise about a month ago that was never very well explained -- possibly it's this? The upshot is that it's totally uninstallable, even using the "put a DOS partition on it" workaround, and modifying the BIOS geometry to match the real geometry (or letting it use its fictional geometry it comes up with). Anyone else have a machine with a WD1007 that they installed instead of upgraded from an existing (non-2.0.5R: working) installation? Terry Lambert terry@cs.weber.edu --- Any opinions in this posting are my own and not those of my present or previous employers.