Eric Whiting writes: > bash$ ls -la > total 18 > drwxrwxr-x 2 hea foundry 109 Oct 30 2001 . > drwxrwxr-x 9 hea foundry 185 Oct 30 2001 .. > -rw-rw-r-- 1 hea foundry 152 Oct 30 2001 > catssubmit.101:10:30:10:48:20: > -rw-rw-r-- 1 hea foundry 2454 Oct 30 2001 mc_5x.cinc > bash$ more mc_5x.cinc > mc_5x.cinc: Is a directory
Is there something in the server logs? Can you try echo 65535 > /proc/sys/sunrpc/nfs_debug on the "penett" just before "cat mc_5x.cinc" and send results (kernel log) to us? Please run cat from non-nfs directory as to avoid garbage in a log. How does mc_5x.cinc looks like on a server? > > NFS server is NFSV3 2.4.13 reiserfs 3.6 (also same error on 2.4.8pre4) > NFS client is solaris8 > > We used to see this sort of error before the knfsd/reiserfs generation > number patch. Now it has been showing up again a little bit. > > Questions: > > 1. IS this a solaris issue? I can go to another identical solaris box > and it works. I reboot the linux nfs server and it still fails. Is > Solaris NFS caching attributes wrong? Most of our sun boxes lack the > latest solaris NFS patch 108727-08, yet our solaris boxes don't have > confusion/trouble with sun or netapps nfs servers like this. If this error survives server reboot it looks like client (i.e., Solaris) problem. > > 2. This solaris box sometimes mounts the linux NFS tree at different > locations (via the automounter) -- /export/home and /export/home/subdir > are both on the same linux filesystem, but will get mounted on the > solaris box as /home/host1 and /home/special1 at the same time. Will > that confuse things on the linux nfs server? > > eric Nikita. >
