[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]
On Thu, Sep 10, 2009 at 5:23 PM, Gordon Henderson wrote: > On Thu, 10 Sep 2009, Martijn Grooten wrote: > >> On Fri, Sep 4, 2009 at 10:59 PM, Gordon Henderson wrote: >>> >>> If you're reading the counters yourself, be aware that they're 32-bit and >>> will wrap at some point, so you need to be able to cater for that. >> >> This isn't really relevant when the scripts compares the values to >> those of a second ago, is it? > > Depends on how fast the interface is going :) > > However 1 second, even at Gbit speeds is OK... (however 40Gb is coming ;-) Well, as the problem was that the speed wasn't what I expected it to be, this wasn't really an issue. :-) >> Am I right to assume that, if the difference in RX bytes (as seen by >> ifconfig) for one network over a given period of time is, say, 1024, >> then 1 kB of traffic has been received by that network connection? In >> particular, does this include traffic that goes through the machine? > > As far as I'm aware, it's on a per interface basis - so it's the interface > that's being monitored, not the "machine". so it's it's a router, then > you'll get 1024 on one interface and another 1024 on another. If it's a > program sucking the data, then you'll still see 1024. Yes, that's kind of what I meant. :-) Anyway, I found iftop to be quite a useful program; it's like top but displays bandwidth usage instead of CPU usage. But in the end it turned out to be a faulty line... Martijn. -- The Mailing List for the Devon & Cornwall LUG http://mailman.dclug.org.uk/listinfo/list FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html