We're mounting a NAS filesystem via NFS to a SPARC Solaris 10 server, and 
had to remount on the fly due to a parameter change on the NAS device.  
The mount -o remount command neither preserved the existing mount 
options, nor read them from /etc/vfstab (they're the same).  AFAIK it 
took all defaults, which caused some errors.  I have resolved the issue 
by executing another remount and explicitly specifying all the mount 
options needed, but this behavior doesn't make sense to me.  Is this 
normal behavior?  I have researched the docs and the Web and don't see 
anything that discusses this.

F. Michael Orr wrote:
> We're mounting a NAS filesystem via NFS to a SPARC Solaris 10 server, and 
> had to remount on the fly due to a parameter change on the NAS device.  
> The mount -o remount command neither preserved the existing mount 
> options, nor read them from /etc/vfstab (they're the same).  AFAIK it 
> took all defaults, which caused some errors.  I have resolved the issue 
> by executing another remount and explicitly specifying all the mount 
> options needed, but this behavior doesn't make sense to me.  Is this 
> normal behavior?  I have researched the docs and the Web and don't see 
> anything that discusses this.

At least this behavior of mount_nfs is consistent with
mount_ufs - it reads /etc/vfstab but then uses its defaults.

Unchanged behavior since Solaris 2.6 (or even before),
so in this sense it is "normal".

-- 
Michael Tosch @ hp : com