[AFS3-std] Request for new fields in rx_statistics

Russ Allbery rra@stanford.edu
Wed, 30 Sep 2009 17:42:23 -0700


Jeffrey Altman <jaltman@secure-endpoints.com> writes:

> As for whether or not "the number of ack packets that have been delayed
> due to excessive aborts" should be sent or not, I wonder if that number
> is really a useful number from the perspective of debugging the server
> state.  From my perspective the question that one is trying to answer is
> whether or not a particular rx_connection or rx_peer is being throttled.
> I don't think reporting the number of delays will help address that
> question.

I just had a file server fail with symptoms that may be related to a
particular client pounding on it and may be related to a problem with the
file server itself.  If I had this statistic available, I could have seen
if the culprit appeared to be user programs trying to access directories
that they aren't permitted to access, or whether something else was
causing the problem.  It would have provided some clarity.

I agree that having this information per-connection is better than having
it per-server, but having *some* information about the frequency of this
problem would be nice.

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