Kurt Buff ha scritto: > All, > > I'm running an ntop installation on a box with the following stats: > > CPU: Intel(R) Celeron(R) CPU 1.70GHz > > real memory = 2138308608 (2039 MB) > avail memory = 2087510016 (1990 MB) > > rl0: <RealTek 8139 10/100BaseTX> <---------------------------- > numbered interface > fxp0: <Intel 82801DB (ICH4) Pro/100 VE Ethernet> > <-------un-numbered, monitoring interface > > I'm running ntop 3.3.1 on FreeBSD 6.2 Stable, and I've placed the > unnumbered interface on a monitor port on an HP switch between the > firewall and our backbone switch. We've just implemented a DS3, and we > had a huge spike in traffic this morning, where the incoming data was > over 20mbits/sec, for about 10 minutes, and then ntop fell over -
Ntop should keep up with throughput, however several reasons could cause a crash (traffic type, size, ntop configuration...). What kind of traffic you had during that 10 mins? > stopped responding to http requests, though it showed as a running > process - and I rebooted the box to bring it back to life. I don't > know that the two are actually related, but it is suspicious, though > the box had been running for several month - the previous reboot was > due to some patching I did. Did you still have ntop logs? > > So, a couple of questions: > > 1) Is it likely that ntop would fall over under that kind of load, or > should it simply have started dropping packets? Read above (configuration issues?) Cheers, Yuri > > 2) Would a beefier box have allowed it to keep up with this load? > > I have a png of the hourly graph showing the gap, if that's of any interest. > > Thanks, > > Kurt > _______________________________________________ > Ntop mailing list > [email protected] > http://listgateway.unipi.it/mailman/listinfo/ntop _______________________________________________ Ntop mailing list [email protected] http://listgateway.unipi.it/mailman/listinfo/ntop
