From owner-freebsd-stable Wed Oct 20 2: 7:32 1999 Delivered-To: freebsd-stable@freebsd.org Received: from metis.salford.ac.uk (metis.salford.ac.uk [146.87.232.15]) by hub.freebsd.org (Postfix) with SMTP id 3CBB914D48 for ; Wed, 20 Oct 1999 02:07:00 -0700 (PDT) (envelope-from M.S.Powell@salford.ac.uk) Received: (qmail 14474 invoked by alias); 20 Oct 1999 09:06:57 -0000 Received: (qmail 14468 invoked from network); 20 Oct 1999 09:06:57 -0000 Received: from plato.salford.ac.uk (146.87.255.76) by metis.salford.ac.uk with SMTP; 20 Oct 1999 09:06:57 -0000 Received: (qmail 67220 invoked by uid 141); 20 Oct 1999 09:06:57 -0000 Date: Wed, 20 Oct 1999 10:06:57 +0100 (BST) From: Mark Powell X-Sender: mark@localhost To: "Joe \"Marcus\" Clarke" Cc: freebsd-stable@freebsd.org Subject: Re: How to fix cant load /kernel - NFS corruption? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Wed, 20 Oct 1999, Joe "Marcus" Clarke wrote: > Are you sure the kernel in root is an ELF kernel? I persoanlly haven't > seen this error, but it looks like the file type for /kernel is not known > to the loader. That's what I thought. However, I've done a straight copy of the whole OS from the other machine. Even trying to boot kernel.GENERIC gives the same error. Okay, tried file and all three kernels in the root are corrupted. They are the correct length, but kernel and kernel.old have a extra 0 byte at the start. kernel.GENERIC has an 0x0a. Everything else is out a byte, so I can only assume what was the last byte is missing. I ftpped the kernel again from the other machine and rebooted. It came up, but I noticed lots of "syntax error: '(' unexpected" during boot. Looks like loads of executables are corrupted. Checked a few data files and some of those are corrupted too. Hmmm, looks like an nfs problem? This ain't supposed to happen? I've done machine copies in the past using tar and rsh, but that doesn't seem possible with the fixit floppy > Can you run ``file /kernel'' on it from the fixit floppy? > Also, according to Kirk's talk at FreeBSDCon, it seems SoftUpdates can (or > will) allow for crashed systems to come back up without needing to run > fsck. Yeah, I've read that document. However, I notice from looking at the source to fsck, that it makes some assumptions about a fs if the soft updates flag is set. I had the soft updates flag set, but didn't actually mount the fs with soft updates on, because I took the option out of the kernel. I had the fs mounted async. I'm wondering if having the flag on was what caused the fs corruption. Although I am aware of the possibilty of fs corruption, just using async. This is the first time it's ever happened though. > I don't think all the support is there yet, but Kirk seems > confident that softupdate is good, and is going to get a lot better. He's > in the process now of trying to sell it to the major industry hitters. > After that, it should become the default in FreeBSD. Well I don't want this to happen again so I'll try soft updates rather than stick with async. Mark Powell - UNIX System Administrator - Clifford Whitworth Building A.I.S., University of Salford, Salford, Manchester, UK. Tel: +44 161 295 5936 Fax: +44 161 295 5888 www.pgp.com for PGP key M.S.Powell@ais.salfrd.ac.uk (spell salford correctly to reply to me) To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message