[OpenAFS] Dummy fileserver putting clients 'on hold'

[Rainer Toebbicke]

Neulinger, Nathan wrote:
> Why are you blocking access to the server during the moves?
> 
> What is that gaining you?
> 

Sorry for putting too many things into the same sentence/paragraph: the 
'moving' was a proposal to solve my original problem that I dismissed as 
impractical.

The original problem was: how can I do a [5-minute] manipulation on a 
fileserver involving disks not being available without upsetting too many users.

The answer is: maneuver the fileserver into a blocking mode for that time, 
such that requests are retried instead of aborted. The interactive users will 
moan and start hitting control-C and the like, but at least the more patient 
ones, batch jobs and running applications will survive.

I'm still not quite happy about the fact that accesses to R/O volumes will 
probably share the same fate instead of going to another replica, but its a 
step forward.


-- 
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Rainer Toebbicke        http://cern.ch/~rtb         rtb@mail.cern.ch  O__
European Laboratory for Particle Physics(CERN) - Geneva, Switzerland   > |
Phone: +41 22 767 8985       Fax: +41 22 767 7155                     ( )\( )