Hi Wai Phang, Counters are built over a period of 5 minutes. So you must also divide by 300 (secs) to get a Mbps value. Thing is by doing that the new value sits around 18-20 Mbps rather than 100Mbps. At this propo I see that your sql_refresh_time value (60) is not aligned with sql_history (5m), which is perfectly legal but clashes with 'sql_dont_try_update: true'. Either go for:
sql_refresh_time: 300 sql_history: 5m sql_dont_try_update: true Or for: sql_refresh_time: 60 sql_history: 5m sql_dont_try_update: false What looks to be happening now is that only the first 60 secs worth of data out of the 300 secs timeslot are inserted in the database and the rest is going lost. And multiplying 18-20Mbps by 300/60 (5) we get back to your expected 100Mbps value. Cheers, Paolo On Wed, Nov 02, 2011 at 01:20:04PM +0800, Wai Phang wrote: > Hi there, > > I am working on pmacct to get the aggregated bandwidth utilization of > different incoming source AS from my switches. > > However, when I compared the aggregated bytes in the sql table against > my snmp, the differences is about 50 over times. > > For example, for a particular AS, I am getting the below, converted to > Mb but from my SNMP RDD, the utilization for that AS if only 100Mb for > that period. > > +---------------------------+------------------------+ > | stamp_inserted | bytes*8/1000000 | > +----------------------------+-----------------------+ > | 2011-11-02 13:05:00 | 5602.0440 | > | 2011-11-02 13:10:00 | 4767.3149 | > +----------------------------+-----------------------+ > > What am I missing here? Thanks! > > Below is my configurations. > > ======================= > daemonize: true > debug: true > plugins: mysql > plugin_pipe_size:10240000 > plugin_buffer_size:10240 > sfacctd_port: 6343 > sfacctd_renormalize: true > aggregate: src_as > sql_user: root > sql_passwd: > sql_table_version: 6 > sql_refresh_time: 60 > sql_history: 5m > sql_history_roundoff: m > logfile: sf.log > sql_dont_try_update: true > ======================= > > _______________________________________________ > pmacct-discussion mailing list > http://www.pmacct.net/#mailinglists _______________________________________________ pmacct-discussion mailing list http://www.pmacct.net/#mailinglists
