ganglia issues
Despite the fact that we do not have a ganglia server running, each instance has a ganglia-client installed which collects stats and sends them nowhere.
We need to disable this ASAP. Which recipe do we have to override for this? Or is this an actual bug in scalarium?
Support Staff 2 Posted by Jonathan Weiss on 02 May, 2012 06:38 PM
The ganglia client should not be running if you don't have a ganglia master. If so this is a bug in Scalarium. Can you please run a configure on an affected instance and paste the link to the log?
3 Posted by till on 01 Jun, 2012 05:56 PM
So I just put up a new instance in our
Playgroundcloud. Ganglia client going to work. Depending on the instance type and e.g. associated volumes, this creates huge load spikes in our monitoring.And regardless if it doesn't – I'd prefer the client to not run.
Support Staff 4 Posted by Jonathan Weiss on 03 Jun, 2012 04:13 PM
Ok, so this is not the Ganglia client running but some cron-jobs gathering metrics for it.
I'll have a look at disabling the cron-jobs if there is no ganglia master.
5 Posted by till on 03 Jun, 2012 04:23 PM
Yeah, sorry – I wasn't be obvious. It's the cronjobs! ;) But ganglia client is installed as well and I am guessing that is where the cronjob comes from.
6 Posted by till on 12 Jun, 2012 04:02 PM
It looks like this happens: the package
ganglia-monitoris installed and it sets up all these cronjobs.Can you guys advice why it installs that when I have no monitoring master?
Support Staff 7 Posted by Jonathan Weiss on 12 Jun, 2012 04:13 PM
We always install the package, we just only start the actual process if you have a master.
What we apparently always do (and shouldn't) is to install the cron-jobs. This is a bug on our side.
8 Posted by till on 11 Jul, 2012 02:22 PM
Do you guys have an update here?
Support Staff 9 Posted by Daniel Hüsch on 12 Jul, 2012 10:45 AM
Unfortunately not yet, but this item is on the agenda.