Issue

Symptom/Problem

• Even if an error occurs, statfs on NFS always returns 0
• When an error like the server down occurs on NFS, if a program calls statfs() system call, it will return 0. However it should return a negative value.

Environment

• Red Hat Enterprise Linux 5.3
• Red Hat Enterprise Linux 4.8

Resolution

• Apply the following errata which addresses this issue. http://rhn.redhat.com/errata/RHSA-2011-0263.html

Root Cause

• nfs_statfs() function shouldn't return 0.
• The original patch upstream.

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;f=fs/nfs/super.c;h=719464a04ddae4fc73dc81ba982ee827631fb9ad;hp=f1eae44b9a1aec5167dfefb934437ccaedba2576;hb=1a0ba9ae485c5fd17d0bff2f14d9dd75b8985593;hpb=d585158b608248a6ba8ae75e234672e048d3fde9

Diagnostic Steps

• Start NFS service:

  # echo "/test        *(rw,no_root_squash)" >> /etc/exports
  # service nfs start
  

• Setup client:

  # mount -t nfs -o soft,timeo=5 server:/test /mnt/nfs
  

• Stop NFS service:

  # service nfs stop
  

• From the client, run program such as:

  #include <stdio.h>
  #include <stdlib.h>
  #include <sys/statvfs.h>
  
  int main()
  {
       struct statvfs sv;
       int rc;
  
       rc = statvfs("/mnt/nfs", &sv);
  
       if (rc < 0) {
            perror("statvfs");
            exit(1);
       }
       return 0;
  }
  

• Waiting  for awhile, it would return with 0 even NFS server is already offline.  The above program returns with 0, which actually should return negative  number according to statfs(2).

• The call trace is as below:

  sys_statfs() => return 0
    vfs_statfs() => return 0
      nfs_statfs() => return 0
        nfs3_proc_statfs() => return -5