innpeer-stats [was: Re: innd logging conventions]

Fabien Tassin fta at sofaraway.org
Thu Dec 28 04:36:34 UTC 2000


According to Matus fantomas Uhlar:
> -> innd(8) records the following convention for innd logging:
> -> 
> 
> Well, I am currently writing some utils that analyze syslog output for my
> employer. I found innd one of programs that are very hard to analyze because
> there's no simple way of analyzing it all, except of playing in patterns or
> substring operations (I am doing it in perl) and thus I will process innd
> later (and maybe make a patch)

I've also written one tool (in C this time) to analyze trafic with peers
because innreport is not good at this job (probably my fault).
I've just made it available there :
  http://www.sofaraway.org/projects/innpeer-stats/

If you try it, you will see that incoming data are underestimated because
of a real lack of checkpoints entries. Depending of the server on the
other side, connections can stay alive more than a day without producing
a single line of log :( I think INN should definitly be modified to
avoid this. (no time yet to produce a patch).

BTW, I'm open to comments on innpeer-stats.

-- 
Fabien Tassin -+- fta at sofaraway.org



More information about the inn-workers mailing list