[rrd-users] Re: stuck building RRD

Craig Packard cpackard at gmail.com
Wed Oct 26 14:57:49 MEST 2005

On 10/25/05, Alex van den Bogaerdt <alex at ergens.op.het.net> wrote:
> On Tue, Oct 25, 2005 at 02:57:21PM -0400, Craig Packard wrote:
> Your heartbeat equals your step size. You must update every 300 seconds,
> 301 seconds is not good enough. Do you really want that? Don't get me
> wrong, this may be OK if you choose to do so. But if you run from some
> kind of schedular, it is very likely that you run this script not at 300
> seconds exactly but rather approximately every 300 seconds. In other
> words,
> 15:10:00, 15:15:01, 15:20:00, 15:25:02 and so on. Each time an interval
> is more than 300 seconds, you have instructed rrdtool to reject the data
> and substitute NaN.

Ok, I think I understand this. Now as a fix I could 'run' the data collector
every 2.5 minutes, that would pretty much make certain that at least one
data point made it in time. I guess I could also raise my heartbeat up to
600 or even 900. My question is if the RRD gets two data points (1 every
2.5min) does it average them together or drop the second one? My
understanding is it will average them since my RRA is AVERAGE but I'm not
sure. I need each data point to be used.

You have xff==0.5 so you need to have two unknown intervals in a row to
> see one blank spot in your data. I think you may need to reconsider
> your feeding script and/or your rrd parameters. Modifying xff would
> only mask the problem (and besides, it should allow more unknowns, not
> less, so you'd need to go to 0.9, not to 0.1)

I'm perplexed how I could be missing two in a row. Does that apply even to a
RRA that is for each step? In other words my RRA isn't average multiple
inputs so I'm thinking if the one it needs isn't there then it is toast. I
could see my incoming data being 'late' but you would think if it was late
it would just treat it like the next 300seconds block (the one it is
currently in) of data.


It has very much, thanks!

> Alex van den Bogaerdt
> http://www.vandenbogaerdt.nl/rrdtool/

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://lists.ee.ethz.ch/rrd-users
WebAdmin    http://lists.ee.ethz.ch/lsg2.cgi

More information about the rrd-users mailing list