Why does ganglia think my host is down?

Posted by NZKoz on Server Fault See other posts from Server Fault or by NZKoz
Published on 2011-03-18T07:51:21Z Indexed on 2011/03/18 8:11 UTC
Read the original article Hit count: 273

Filed under:
|
|

I have ganglia set up to monitor our staging server, it's working great but I'm confused by the definition of 'down' to ganglia.

There's a single node, running gmetad, gmond and the web frontend, but some small percentage of the time the web frontend shows some confusing output. Despite the fact that it's a single server in the cluster, and that server is the one serving the web interface, the dashboard output insists that the host is down. Then below that it has a graph which shows 50% down, 50% up. You can see an example of this here: http://i.imgur.com/MCWaS.jpg

There's obviously something confusing ganglia somewhere, but I'm not sure where to start looking. Unfortunately googling for any combination of 'ganglia' 'down' 'metric name' seems return nothing but other people's ganglia installations displaying the same nonsense. Any tips on where to start looking would be greatly appreciated

© Server Fault or respective owner

Related posts about linux

Related posts about monitoring