collectd and slave/master usage

[HalfEatenPie]

Seeming as we have a ton of incredibly knowledgeable people around here, I have a question to make.

I have Collectd installed on both the slave (monitored) server and the master (monitoring) server.  Unfortunately, the slave server is a very low-resource server (64 MB RAM) and therefore doesn't have a whole lot of hard drive space.  How can I setup collectd to delete/remove the local copy of the data so that it doesn't increase the amount of storage space used but still not affect the data on the master monitoring node?  

In other words, instead of having two copies of the data (one on the slave server and another on the master server), how can I set it up to only have the data on the master server? 

Thanks!

Edit: This is my projected storage usage thanks to Collectd (on the master server) with a trendline (in red).  



Currently, my projections state I'll be needing more space around February.  So while this isn't the slave (monitored) node's usage, you can see how it can be a problem later on down the road.  I want to look into minimizing the storage impact without losing too much of the resolution of the data (in addition to centralizing the data to one server instead of two).  

 

[wlanboy]

Can't believe that collectd does not have a call for that.

So you can do it the dirty way:

service collectd stop && find /var/db/collectd/rrd/localhost/* -type d | xargs rm -Rf && rm /data/rrd_dir.tar.bz2 && service collectd start

 

[HalfEatenPie]

Hm.  I guess that probably is the easiest way...

Should probably have it set on a cron for every other week haha.