From owner-freebsd-bugs Fri Jul 31 07:25:37 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id HAA01228 for freebsd-bugs-outgoing; Fri, 31 Jul 1998 07:25:37 -0700 (PDT) (envelope-from owner-freebsd-bugs@FreeBSD.ORG) Received: from elvis.vnet.net (elvis.vnet.net [166.82.1.5]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id HAA01223 for ; Fri, 31 Jul 1998 07:25:35 -0700 (PDT) (envelope-from rivers@dignus.com) Received: from dignus.com (ponds.vnet.net [166.82.177.48]) by elvis.vnet.net (8.8.8/8.8.4) with ESMTP id KAA23258; Fri, 31 Jul 1998 10:25:28 -0400 (EDT) Received: from lakes.dignus.com (lakes [10.0.0.3]) by dignus.com (8.8.8/8.8.5) with ESMTP id KAA02296; Fri, 31 Jul 1998 10:57:57 -0400 (EDT) Received: (from rivers@localhost) by lakes.dignus.com (8.8.8/8.6.9) id KAA05492; Fri, 31 Jul 1998 10:29:32 -0400 (EDT) Date: Fri, 31 Jul 1998 10:29:32 -0400 (EDT) From: Thomas David Rivers Message-Id: <199807311429.KAA05492@lakes.dignus.com> To: rivers@dignus.com, walter@fortean.com Subject: Re: kern/6102 Cc: dillon@best.net, dima@best.net, freebsd-bugs@FreeBSD.ORG, newmike@teleport.com In-Reply-To: Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > In my "daily panic" days; this is one of the panics I got every now-and-then, > > although not nearly as frequently as the "dup alloc" one. > > David, > > As a long-time list lurker, I remember the 'David Rivers Memorial Panic' > messages... IIRC, they had to do with a heavily loaded news server, no? > What I can't remember is what controller the disk subsystem was using. > > Cheers, > > - Bruce > > PS: Wasn't it all ocurring on a 386 system too? > Well, the answer to your question is "yes" and "no" :-) Yes, I got the problems in a 386 with an Adaptec 1542B... but, I also reproduced the problem on a 486 with an IDE adapter... Unfortunately, due to the general situation of Data General here in RTP, NC - I lost my full news feed; and so can no longer reproduce the panics with a heavily loaded news server... [the panic almost always occurred when Cnews was running the nightly expiration...] But, I believe I had narrowed the problem down to a issue which I was able to reproduce on a stand-alone machine.... write 0xff to a disk partition, run newfs, then run fsck on that newly created file system. On my reproduction machine (another 386 with a 1542B), the fsck will fail because some of the 0x00's which should have been written to the disk didn't make it... I can make that machine available over the net, if anyone wants to try it out, build kernels and debug the problem... The machine is set up to boot from a floppy with a serial console. So, you can remotely install a new kernel, reboot the machine and do kernel debugging, etc... [It's off-the-net right now, but a quick trip to Radio Shack for some RG58 cable will get it back :-) ] - Dave Rivers - To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message