Dunc wrote: > It runs on Linux fine, so that's probably all you need to know. It > shouldn't matter if it's virtualised or not. > I'm definitely running it on a Linux guest inside Xen if that helps.
> Johnston, Robert Bruce wrote: > > I’m new to MRTG tools and have just been asked about the > possibility > > of running them under Linux on Virtual HyperV. Has anyone > experienced > > any Success or Failures using this configuration? You can run MRTG under a fully virtualised environment; however you need to be aware of the issues involving "Clock Skew". Note that Paravirtualised environments (available in Xen and VMWare4) do not suffer from this problem as they use the system hardware clock, not a virtual clock. If there is contention in the virtual environment, or if the virtual machine 'sleeps' for any time, the then virtual clock can start to lag behind the real clock and then skips forward quickly to catch up. Lag can be up to a minute in some environments, although it is more likely to be in the region of 10s. This can mess up any time-based calculations and cause inaccuracy; for MRTG (which polls every 5min) the effect will be relatively small but it can be more serious on things like OS-reported CPU usage which is calculated over a shorter interval, or if you are using a data-collection plugin which calculates rates over a shorter interval. There's a section on this in the book http://www.steveshipway.org/book/ (obligatory plug). Steve _______________________________________________ mrtg mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/mrtg
