[OpenAFS] Fail over to replica sites

[Russ Allbery]

Jimmy Engelbrecht <jimmy@e.kth.se> writes:
> Russ Allbery <rra@stanford.edu> writes:

>> The server that went down is in our server room on one side of campus,
>> and the clients that I was watching in particular are in the server
>> room and student cluster on the other side of campus, in several
>> different subnets, all different than the subnet the server was on.
>> (And all switched.)

>> Not all the clients had problems, but all of the volumes in question
>> are replicated on three sites, so I don't know if all the clients that
>> were bound to that particular server had problems or not.

> Did you ever run "fs checks" on a mashine that was unhappy ?

I'm not sure if we did.  That sometimes does kick it out of that state
when I've seen it in that state before.

Oh, and all of the systems in question were running OpenAFS 1.2.5 as the
client and were running various versions of Solaris between 2.6 and 8,
mostly Solaris 8.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>