Some NFS file transfers fail and hang automounting

From Linux NFS

About

• [Bug 126213]
• Reported by: Philippe Dax (June 15, 2010)
• Fixed by: Trond Myklebust (June 16, 2010)
• Kernel version: 2.6.33.5-112.fc13.x86_64

Symptoms

• Given a file "foo" of 50Mb on a remote machine "remote".
  • This command will never finish

    <localmachine $> cp /remote_mount_point/foo bar 

  • bar will have a size less than foo.
  • automounting of the local machine is hung.

• The following message will show up in /var/log/messages

kernel: Callback slot table overflowed

• The problem doesn't occur if foo has a size less than 10Mb
• The final size of bar appears to be random

• This incident occurs with:

  sunrpc.tcp_slot_table_entries = 16

• This incident does NOT occur with:

  sunrpc.tcp_slot_table_entries = 32

Resolution

This problem was fixed by:

commit b76ce56192bcf618013fb9aecd83488cffd645cc
Author: Trond Myklebust <Trond.Myklebust@netapp.com>
Date:   Wed Jun 16 13:57:32 2010 -0400

    SUNRPC: Fix a re-entrancy bug in xs_tcp_read_calldir()
    
    If the attempt to read the calldir fails, then instead of storing the read
    bytes, we currently discard them. This leads to a garbage final result when
    upon re-entry to the same routine, we read the remaining bytes.
    
    Fixes the regression in bugzilla number 16213. Please see
        https://bugzilla.kernel.org/show_bug.cgi?id=16213
    
    Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
    Cc: stable@kernel.org