|
At 09:04 AM 6/4/2002 +0800, you wrote:
>>>ipvsadm -L -n --rate
>>
>>This call does not seem getting reliable rate returns. Is that
>>a problem other also see or just my testing setup?
>
>During my tests, I recorded the rate periodically (from
>/proc/net/ip_vs_stats), and the results I obtained are very close to what the
>benchmarking client had measured as the overall traffic figure. I assume the
>rates given by ipvsadm are the same as the ones in the proc fs.
>
>Make sure you zero the ipvs counters before you start measuring (ipvsadm -Z).
>
>What makes you think the results are unreliable?
>
>
>Jeremy
Thanks for the reply. I found out the problem was one of our
little test program sent a wrong parameter to the setsocketopt()
call that ipvsadm also called. It seems the wrong
parameter cause some confusion in the setsocketopt() inside
the kernel, and that is remembered until next time system
reboot :<, so that other program could not get correct result,
either. If we did not run that little test program of ours since
reboot, ipvsadm will keep working fine. Sorry for the false alarm.
>_______________________________________________
>LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
>Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
>or go to http://www.in-addr.de/mailman/listinfo/lvs-users
|
