Date: Thu, 29 Nov 2001 10:37:36 -0500 (EST) From: Andrew Gallatin <gallatin@cs.duke.edu> To: Paul Herman <pherman@frenchfries.net> Cc: freebsd-alpha@freebsd.org Subject: Re: NFS Kernel Panic. Unaligned Access? Message-ID: <15366.22080.471677.562171@grasshopper.cs.duke.edu> In-Reply-To: <200111290217.fAT2HZQ01665@tick.sc.omation.com> References: <15365.32857.392495.737510@grasshopper.cs.duke.edu> <200111290217.fAT2HZQ01665@tick.sc.omation.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Paul Herman writes: > On Wed, 28 Nov 2001, Andrew Gallatin wrote: > > > Can you characterize your environment a little more? > > Hi Drew, > > Thanks for the in-depth explanation. > > A little about the environment: the client is 4.4-RELEASE alpha and > the server is 4.3-RELEASE i386 both with NFS compiled in kernel. > The mount point using NFSv3 over TCP: > > tick$ df /usr/home > Filesystem 1K-blocks Used Avail Capacity Mounted on > fs2:/u01/home 551540862 467896476 39521118 92% /usr/home > > At the time it was crashing, it was a particular file. Every time > after reboot I tried to access it, it panic'ed. I went to lunch > this afternoot, came back, and since then I couldn't reproduce it. > I will try again tomorrow, and if I do I'll post a tcpdump. The more that I think about it, the less I think that a TCP dump would help -- the client is forming a read rpc, so it hasn't talked to the server yet for this particular transaction. If it keeps happening, can you switch to UDP & see if it goes away? Its a lot faster in most cases, and since I've never seen this with 10+ alphas over 2 years running NFS over UDP, I'm figuring it might be safer.. Drew To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?15366.22080.471677.562171>