[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

How to debug something like this?



Hi all,

this maybe offtopic, but I am almost desparate.

Last night I migrated the nedbsd.nl 3.5 server to a other machine running 3.7. 
The specs are the same, except for the network card. On 3.5 I used
fxp0, now I use xl0.

Everything works fine (well, I do miss nullfs but use loopback nfs as
a substitute now), except for the jabber services.
I use jabberd 1.4.3.1 and switched the aim/icq/msn transports to
python based transports.
Jabberd itself is compiled exactly the same way and uses the same
configuration. For some reason jabberd sometimes eats 50% CPU and the
loadavg goes up to 16.

ktrace did not reveal something usefull, systat and top did not tell
me much either.
How can I find what is causing this?
It really hurts my server performance every 15 minutes or so, and my
statistics aren't very good either:
http://public.jabbernet.dk/mrtg/nedbsd.nl.html
I am obviously unreachable sometimes.

Regards,
Wijnand