Hello,
I've got linux boxes working as firewalls (netfilter) and router (policy
routing, QoS, etc) on my core network, for some time they have been doing
just fine. But now that the traffic is quite high already, and I'm
upgrading the main link to internet, I would like to get some statistics
on how the machines are doing. Is there any way to tell when a machine is
starting to reach it's limits and run out of CPU time. As far as I've
understund, the packet filtering, routing and queuing code, etc is run by
no spesific task and so by is counted to the CPU usage of idle task.
So, if there is no way to log this now, would it be possible solve this,
start using own thread for this kind of kernel activities, move the
network code as own kernel thread or count the time used by interrupts,
or something. CPU time of idle task makes little sense if the main service
is counted as idle. It would be possible to estimate the CPU usage of
this functionality, because knowing when you're running out of resources
if vital.
If there is a way to account this, please, let me know!
- Sampsa Ranta
sampsa@netsonic.fi
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Jun 15 2000 - 21:00:22 EST