seanbru at yahoo-inc.com
Tue Aug 31 19:51:45 UTC 2010
On Tue, 2010-08-31 at 12:27 -0700, Garrett Cooper wrote:
> On Tue, Aug 31, 2010 at 11:32 AM, Sean Bruno <seanbru at yahoo-inc.com> wrote:
> > On Tue, 2010-08-31 at 09:29 -0700, N.J. Mann wrote:
> >> In message <1283271393.3665.13.camel at home-yahoo>,
> >> Sean Bruno (seanbru at yahoo-inc.com) wrote:
> >> > An issue in testing came up where we were trying to mount NFS
> >> > directories from the same server, i.e. a machine trying to mount an NFS
> >> > dir on itself.
> >> >
> >> > Because mountcritremote runs before the NFS server is up, we modified
> >> > the REQUIRES section and appended nfsd.
> >> Is this not what the 'late' attribute in /etc/fstab is for?
> >> e.g.
> >> oberon% grep /usr/ports /etc/fstab
> >> oberon:/export/usr/ports /usr/ports nfs late,rw,tcp
> Specifying `late' seems extremely unnecessary for NFS mounts. If
> mountcritremote can't properly depend on nmbd, nfsd, smbd for mounting
> filesystems, then it seems like a rc script dependency issue
> (otherwise, mountcritremote is basically mountcritlocal, waiting for
> some networking connectivity to come up, right?).
Yup. That's my impression. In my user's application, they have 100's
of NFS mounts, of which ONE is a local-to-the-server mount. So, in this
case, I don't really want to delay the mounting,
I want the mounting to succeed at the time mountcritremote is invoked.
The user application is doing something silly and severely knuckle
headed, but it seems it should work in the way I propose instead of
mounting late in "mountlate"
However, I'm pretty ignorant of the dependencies in the system of rc
scripts. There may be dragons here.
More information about the freebsd-hackers