[OpenAFS] client doesn't see newly created replication sites

[Ken Raeburn]

> Excerpts from mail: 16-Nov-01 Re: [OpenAFS] client doesn'.. Charles
> Clancy@xauth.net (1636*)
> 
> > The solution is to mount the '.readonly' volume instead of just the '#'
> > volume.  This guarantees you hit the RO one.
> 
> > The other pseudo-solution I've seen is to reboot the machine.  Then it
> > will rebuild the cache, and with 3/4 probability decide to use one of the
> > RO volumes instead.

That sounds rather broken, and not what I've been told about AFS
handling of mount points.  As I understand it, going through a '#'
mount point in a readonly volume gets you the RO volume if it exists,
or the RW one if the RO one doesn't exist.

Craig_Everhart@transarc.com writes:
> The solution is neither of these.  The normal mount point is fine.  What
> you have to do is re-traverse the mount point, so that you get into the
> replicated volume instead of the RW.  Your current working directory
> isn't supposed to change out from under you.

You mean just "cd" to that pathname again?  That was the first thing I
thought of; I tried it by itself, and after every attempt to find a
way to get the kernel to look up the volume information again
(including traversing a mount point explicitly for the .readonly
volume) or discard its previously cached informatio.

FWIW, my machine wedged up and had to be rebooted, and I'm not having
the problem any more.  But it still leaves me wondering if I'm going
to mess things up for other OpenAFS people on campus who might still
be using the RW volume if I update some of the software...

Ken