[OpenAFS] Re: Best way to debug "Lost contact with file server"

Russ Allbery rra@stanford.edu
Tue, 28 Feb 2012 16:59:55 -0800

Benjamin Kaduk <kaduk@MIT.EDU> writes:

> Er, I would have thought that this would cause a group of clients
> deployed near each other to all prefer the same replica, which is not
> really "random".  That is, though from the point of view of a single
> client the choice may be random, from the point of view of managing
> server load it is not.  Is that incorrect?

No, if you happen to be in historic class A space and have some servers in
the same class A and others not, or in historic class B space and some in
the same class B and others not, or historic class C space, etc., then the
algorithm will consider the ones in your same class to be much closer in
ways that you may not be horribly happy with.

But it's all based on the historic A/B/C ranges, I believe.

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