[rrd-users] rrdupdate behaviour when data already exist (was: Tips and thoughts on how to optimally import data points from MRTG .log files into new RRD databases)

[Simon Hobson]

Cyrille Bollu wrote:

>Now, I have one additionnal question:
>
>For a few months, we are also collecting data from these sensors via 
>another monitoring system using rrdtools.
>
>So, I would like to import data from this other monitoring system 
>into the .RRD that I will generate from MRTG (ie: using "rrdtools 
>update").
>
>But, since there will already be some data, I'm wondering what will 
>be the behaviour of rrdtools in this case (ie: when one inserts 
>values at times when there are already data).

Simple, it will reject the updates, more specifically, you can only 
enter data with ascending timestamps. So you can't insert old data 
into an existing rrd file, once the time has passed then that's it.

If you want to combine data from several sources, over different time 
periods, then you have two options :

1) Build your own rrd file, either directly (sound hard to me !), or 
by building an XML file you can import.

or

2) Generate a list of timestamp/data sets where the timestamp is 
always increasing between samples. You can then feed these into the 
rrd as multiple updates and let the rrd tools take care of the 
consolidation etc as they go.