[OpenAFS] How to stop the access to an AFS volume ?

[Jeffrey Hutzelman]

On Thursday, October 28, 2004 08:54:51 +0200 giovanni bracco 
<bracco@frascati.enea.it> wrote:

> On the test cell have selected the machine which is the sync site and I
> have  run on that a script which does the following:
> 1) stop the vlserver
> 2) apply the patch on vldb.DB0
> 3) restart the vlserver
> All that takes just a few seconds, so soon after the restart the same
> dbserver  is again the sync site for vldb.

OK.  That will work as long as you keep the server down for a short enough 
time that it is not possible for another server to win the coordinator 
election.  If another site becomes coordinator, it will relabel the 
database and your change will not take effect even if there were no changes 
to the contents of the VLDB.




> I have also noticed, it may be obvious of course, that just
> changing the  MaxVolumeId in the vldb.DB0 does not triggers the update on
> the other  dbservers. The update happens only when a new volume is
> created.

Yes; the update will happen only when the database version changes.  You 
can insure that the update will happen right away by bumping the database 
version at the same time you make the other change.  The version consists 
of an epoch and sequence number, and is part of the data stored in 
vldb.DBSYS1.


> The option (2) seems more attractive to me and I think it can be a very
> usefull utility - it is exactely what I was looking for - so if you can
> find  some time I am very interested in having that!

No problem; I'll try to do that shortly.

-- Jeffrey T. Hutzelman (N3NHS) <jhutz+@cmu.edu>
   Sr. Research Systems Programmer
   School of Computer Science - Research Computing Facility
   Carnegie Mellon University - Pittsburgh, PA