[rrd-users] Re: RRDtool "GMT-centric" problem?

Alex van den Bogaerdt alex at slot.hollandcasino.nl
Sun Jan 16 23:48:45 MET 2000

steve rader wrote:
>                  I've found that the heartbeat of exactly one day
> can generate NaN for a full day if the updates are 1day+2sec or
> more apart.  So the best rrdtool create would be something like:
>  rrdtool create page.rrd --start 946788800 --step 3600 \
>  DS:pages:GAUGE:86520:0:1000 \
>  RRA:AVERAGE:0.958:1:240
> which allows for updates up to 1day+2min (86520) apart.

This would lead you back to the previous problem.  If you update the
database on times that are not *exactly* on the boundaries that rrdtool
keeps, you won't find the entered values in the RRD.

Consider intervals I1 and I2.  Interval I1 starts on boundary B1 and
ends on boundary B2.  I2 goes from B2 to B3.

You update the database at B1, B2 + 1 minute and at B2.  Interval I1
will have the value you've entered at time B2 + 1 minute (assuming it is
a gauge value).  Interval I2 is also influenced by this update.  The
remaining part of interval I2 is determined by the value you enter when
updating at time B3.

Interval 2 will show a value that is mostly the value at time B3 but
has a part of the value from the update at time B2 in it.
It will be (60/86400)*V2 + ((86400-60)/86400)*V3.

Given the following input:

   rrdtool update page.rrd \
      946789200:75 \
      946875600:2 \
      946962060:31 \
      947048400:8 \

it will be approximately 8.0159722204 which is rounded to 8.02 ...
In this example:
B1 == 946875600
B2 == 946962000
B3 == 947048400
V1 == 2
V2 == 31

If you want to see the values you enter, you must use the correct times.
Should RRDtool do it otherwise, it would not represent the true situation.
It is showing you the value from B2 to B3, and you entered two different
readings during that interval.

 / alex at slot.hollandcasino.nl                  alex at ergens.op.het.net \
| work                                                         private |
| My employer is capable of speaking therefore I speak only for myself |

Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:rrd-users-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/rrd-users

More information about the rrd-users mailing list