Problem using Graphite with RRDCACHED

Asked by Maziyar Mirabedini

Hi there,

I have Ganglia 3.5 working with RRDTool 1.4.7 and RRDCACHED perfectly. I also have Graphite 0.9.10 installed working great.
In the past I used symbolic links to make RRDs available through Graphite. But I heard about how using FLUSHRRDCACHED would enable you to do the same (Not sure if that is accurate).

I'm trying to use FLUSHRRDCACHED as outlined in the link below to be able to use Graphite to draw the graphs.

https://code.launchpad.net/~shufgy/graphite/graphite-rrdcached/+merge/100973

After configuring the FLUSHRRDCACHED in webapp/local_settings.py I'm still unable to see the metrics anywhere. I checked the logs under /opt/graphite/storage/logs and don't see anything that would be related. This is a brand new test environment and I don't have symblic links to the RRDs set up.

What is this parameter really supposed to do? Am I supposed to be able to browse to different metrics through the Graphite Web UI?

Thanks for the help!

Question information

Language:
English Edit question
Status:
Solved
For:
Graphite Edit question
Assignee:
No assignee Edit question
Solved by:
Maziyar Mirabedini
Solved:
Last query:
Last reply:
Revision history for this message
Maziyar Mirabedini (maziyar-mirabedini) said :
#1

Any help on this is really appreciated!

Revision history for this message
Maziyar Mirabedini (maziyar-mirabedini) said :
#2

Figure this one out...

For those who struggled with this..I'm including some notes that hopefully will help. The problem here is lack of documentation on Ganglia.

As of 3.3.x Ganglia supports sending the metrics directly to Carbon without any plugin.
Here's the thread: http://<email address hidden>/msg06950.html

In order to do that you will need to add the following to Gmetad configuration file. Note that default port is 2003 so it can be omitted.

carbon_server <host> or <ip>
#carbon_port "2003"
graphite_prefix "ganglia"

The parameter "FLUSHRRDCACHED" is only added to accommodate those who use RRDCACHED for i/o reasons. I'll document all the steps on my blog when I have a chance.