Date: Tue, 19 Nov 1996 15:11:08 +1030 (CST) From: Michael Smith <msmith@atrad.adelaide.edu.au> To: ccsanady@friley216.res.iastate.edu (Chris Csanady) Cc: msmith@atrad.adelaide.edu.au, hackers@freebsd.org Subject: Re: vnode_pager_input: I/O read error (2.1.5) Message-ID: <199611190441.PAA27023@genesis.atrad.adelaide.edu.au> In-Reply-To: <199611190356.VAA07278@friley216.res.iastate.edu> from Chris Csanady at "Nov 18, 96 09:56:14 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
Chris Csanady stands accused of saying: > > Swap is local, while all binaries and home dirs are mounted over NFS > from an Irix 6.2 server. I've seen this error from a 2.1.5 server > as well though. Our network is fast and very rarely saturated, so I > do not believe that it would actually lose contact with the server > unless its a server problem. However the same thing works from our > SGI cluster, so I would not think that the case. We will be > installing patches this friday, one of which i think is NFS related, > so I will check that again though. Ok; I've seen similar things with NFS as well, although normally caused by cable dropouts. (In our workshop we use coax because it's more suited to rapid teardown and rearrangement and RJ45's wouldn't last long, but with the amount of coax strung between racks it's easy to get the wrong one...) Which NFS mount options are you using? It's possible that the NFS code is giving up very quickly and not retrying the read, so even light network traffic _could_ result in a lost/damaged datagram. Do the machines show any/many input errors for the network interface? -- ]] Mike Smith, Software Engineer msmith@gsoft.com.au [[ ]] Genesis Software genesis@gsoft.com.au [[ ]] High-speed data acquisition and (GSM mobile) 0411-222-496 [[ ]] realtime instrument control. (ph) +61-8-8267-3493 [[ ]] Unix hardware collector. "Where are your PEZ?" The Tick [[
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199611190441.PAA27023>