We're experiencing problems with some legacy software when it comes to NFS
access. Even though files are visible in a terminal and can be accessed with
standard shell tools and vi, this software typically complains that the files
are empty or not syntactically correct.
The NFS filesystems in question are 8TB+ XFS filesystems mounted with
"delaylog,inode64,logbsize=32k,logdev=/dev/sda2,nobarrier,quota" options,
and I suspect that inode64 may have to do with the observed behaviour. The
server is running CentOS 6.3 + all patches.
The clients exhibiting the problem are running CentOS 5.4 and CentOS 5.8
x84_64. Interesting enough, the application (which is available in 32-bit
only) appears to work fine on RHEL3 32-bit; we have no CentOS5 32-bit
systems to test.
I was wondering if anyone has seen such behaviour and whether it can be
corrected using nfs mount options. The only other solution I have at the
moment is reformatting to ext4, i.e. ugly.
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
