Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 10 Dec 1996 17:22:49 +0200 (EET)
From:      Marko Lamminen <hawk@nixu.fi>
To:        freebsd-current@freebsd.org
Subject:   Re: NFS problems with 2.2-current ?
Message-ID:  <ML-2.2.850231369.5627.hawk@flauta.nixu.fi>
In-Reply-To: <m0vXSma-00037EC@viking.ucsalf.ac.uk>

next in thread | previous in thread | raw e-mail | index | archive | help

  I have no idea if this is related but I'll post this anyway since someone
  might be having same problems.

  Yesterday I posted about a make reinstall that had gone awry. I spent
  many hours debugging this a found the the most probably cause, NFS. 

  Situation was this:

   I had a pentium pro which I used to compile and install a 3.0-current,
   everything went fine (execpt it doesn't boot from disk bootblock). After
   testing it a bit I decided to install 3.0-current to my other machine 
   (an older 486) using the 'reinstall' option and NFS mounted src and obj
   trees. The make reinstall went _almost_ smoothly (there was trouble but
   as it fixed by just doing it again I suspected some older program inferring
   with a new) and after installing a new kernel it was time to reboot. Now
   there were problems, it was unable to mount root writable (mount -o update
   /) because mount core dumped (bus error). Various other binaries were also
   behaving strangly (VI coredumped if I pressed ESC). After extensive testing
   of the new kernel (failsafe,diagnostics,ddb,etc.etc.), I was about ready to
   give up but happened to take a cksum of 'mount' and realized that altough
   the mount was same size as the original (in the pentium pro) it had 
   different checksum. After goin through all of the system files with md5
   I had found over 500 defected files (out of 5500 total) and used rcp to
   copy those from the pentium pro. After that things have been working 
   propely.

   I suspect that it was the NFS that had defected these files but I haven't
   had time to do extensive testing which I will do when get the time. Of
   course the problem might also be the int the harddisk but they would have
   to be quite subtle for the driver not to notice them and the disk has so far
   been quite reliable (4 gig Seagate Hawk4&1542 adaptec).

   I remember hearing something about nfs defecting files sometime agou but
   I don't quite remember what was it that caused it.

 - Marko



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?ML-2.2.850231369.5627.hawk>