
That FAQ is located at 

   http://humdi.net/vnstat/FAQ 

since that's the easiest way to keep it updated.


----


A snapshot of the FAQ (updated 30.12.2007):


Is there some kind of support forum available?

  The forum is located at http://forums.humdi.net/ and is back in use again
  as of 18.12.2007 including old messages. User accounts have however been
  wiped. The forum was previously closed in summer 2006 after cleaning spam
  messages started to take much more time than replying to actual user posts.


How should dialup users use vnStat? There's some errors because the interface
isn't available when offline.

  That's all explained at the end of the README. The idea is to include
  vnStat with enable/disable parameters in scripts related with the used
  interface. Example scripts can be found from the pppd directory that came
  with the source package.


Does vnStat work for PPPoE users?

  Yes. PPPoE is basically like a dialup connection and it can be monitored
  in the same way as any ppp interface. However, with PPPoE it's usually also
  possible to bring up the LAN interface used for the connection without
  configuring any ip to it. Monitoring that interface is a good alternative
  since it doesn't go down and up when the connection is closed and opened
  again.


How do I stop monitoring an interface?

  Go to the vnStat database directory (default: /var/lib/vnstat) and delete
  the database file named according to the interface you want to stop
  monitoring.


How do I uninstall vnStat?

  You only need to run 'make uninstall' in the directory that comes when the
  .tar.gz is extracted. Just make sure it's the same version you have
  installed. If you've used a binary package included with the distribution
  then refer to intructions provided by the package manager.


Is the MB value reported by vnStat 10^6 or 2^20 bytes?

  2^20 bytes. It could as well be MiB but due to space limitations in some
  outputs that isn't used.


Why kB and not KB?

  Because it's kilo (k) and not kelvin (K) bytes.


What does the 'estimated' value mean?

  The estimated value is an calculated average that tries to predict the total
  traffic for the current day/month based on previous traffic. This estimate
  works fairly well if the monitored interface has constant traffic of the same
  magnitude (like game servers). It can't predic peaks but the accuracy
  usually gets better by the end of the day/month.


Why isn't the estimated value shown with --dumpdb?

  That features only dumps the database and since the estimate is always
  calculated in real time, there's no reason to write it into the database.


How is the estimated value calculated?

  estimate = ( x / y ) * z

    x = traffic so far for the day/month
    y = passed minutes/hours ...
    z = total minutes/hours ...


Why does vnStat show sometimes multiple GB of traffic although my network
connection can't even transfer that much?
     OR
Each update adds the complete amount of traffic and 'vnstat -u -D' claims that
the system has been rebooted between updates.

  That's most likely a broken kernel. Run with --testkernel (version 1.2+ 
  required) and see the result. See the next question for some explanation.


Why is there problems with most 2.4.18 kernels?

  Every current kernel has a btime (boot time) value in /proc/stat that
  indicates when the system was booted. vnStat uses this to know when the
  interface counters in the database should be reseted. Unfortunately, some
  kernels don't keep the btime value static even when the system isn't
  rebooted.


About bug reports

  Any bug report should at least include an explanation about how the bug can
  be reproduced. Having output dumps usually helps and the --dumpdb feature
  should be used if there's some reason to assume the bug has something to do
  with the database. Also include information about the used distribution,
  kernel (uname -a), compiler (gcc --version) and network interface card.
  Read the report again before sending it. :)
