Date: Sat, 5 May 2001 10:33:53 +0000 (GMT+00:00) From: auto228677@hushmail.com To: freebsd-questions@freebsd.org Subject: FreeBSD 4.2 kernel panics during installation (ffs_clusteralloc) Message-ID: <200105051033.DAA10093@user8.hushmail.com>
next in thread | raw e-mail | index | archive | help
--Hushpart_boundary_gOmSvITgyUDeHuylZEcQEZhwlGHBxaTE Content-type: text/plain Hello, I've been using FreeBSD 3.3-RELEASE for over a year now and decided to perform a fresh install of FreeBSD 4.2-RELEASE. I bought the official 4 CD set as I had done for 3.3. I boot from my CD-ROM drive, eliminate unneeded device drivers, allow FreeBSD to use my entire disk, allow the install program to choose its default filesystem sizes, select the X-User canned set, and then sit back and wait for the distribution files to be installed. Much to my dismay, the kernel panics 9 times out of 10 while performing the 'bin' install. The message is as follows: Panic: ffs_clusteralloc: map mismatch My machine then reboots. If, however, it manages to install 'bin' with some apparent effort, it will choke soon after with an identical message in one of the next installation stages. I have not been able to get a working system. Hitting CTRL+F2 reveals numerous instances of the following message: ad0: READ command timeout tag=0 serv=0 - resetting ata0: resetting devices .. done So then I thought the smart thing to do would be to install only the 'bin' files, just to get a working system (the rest could be installed at a later time with sysinstall). After enduring half a dozen reboots, it finally managed to trudge through to 100% of the 'bin' files. What then? A nice dialog box pops up saying the following: Can't copy /kernel Ugh! Ok, so I reinstalled FreeBSD 3.3 (which encounters no problems at all -- wd0 driver) and fired up a web browser. Some searching on google.com brought me to the FreeBSD bug reports page. A particular bug report caught my attention: http://www.FreeBSD.org/cgi/query-pr.cgi?pr=16740 The kernel panics with "ffs_clusteralloc: map mismatch" This "critical" problem has been unresolved for almost a year. One of the contributors to the PR (Chris Hardie) narrowed the problem down to heavy disk activity. This would make sense in my case, because of the heavy disk activity during the installation procedure. The fact that people are experiencing this same problem on different disk devices (both SCSI and IDE) leads me to conclude it must be some kernel bug. Fixes? The PR mentions that sysctl -w vfs.ffs.doreallocblks=0 may provide a workaround for the problem by avoiding a certain code path, but some mentioned that this introduces a different bug. In any event, how would I issue that command during the 4.2 install? And what if it fails to do anything? Any help is much appreciated. I'm no kernel hacker, so please dumb things down a bit :-) Thanks in advance. Hardware ======== Intel Pentium 200MMX 64MB RAM 2.5GB HDD (IDE) The hardware is in excellent condition. Free, encrypted, secure Web-based email at www.hushmail.com --Hushpart_boundary_gOmSvITgyUDeHuylZEcQEZhwlGHBxaTE-- IMPORTANT NOTICE: If you are not using HushMail, this message could have been read easily by the many people who have access to your open personal email messages. Get your FREE, totally secure email address at http://www.hushmail.com. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200105051033.DAA10093>