[okfn-help] Moved munin master and munin monitoring working on all nodes

Sun Aug 29 11:21:52 BST 2010

Rufus Pollock wrote:
> Yesterday with Martin Keegan's help I moved the munin master from us1
> to eu1 in order to:
> 
> a) improve the munin monitoring (intermittency of graphs was being
> caused by poor performance on us1)
> b) to pave the way for reducing load on us1 by removing munin monitoring
> 
> As of yesterday evening:
> 
>   <http://munin.okfn.org/> is now running from the new master
>   <http://munin.us1.okfn.org> shows the munin master on us1
> 
> There was some break in monitoring as it turned out hte AWS firewall
> needed to be updated to let eu1 monitor AWS hosts. After allowing a
> few days to check all is ok I propose we turn off the master on us1.
> 
> I have documented some of the lessons:
> 
> <http://knowledgeforge.net/okfn/tasks/wiki/MonitoringService>
> 
> There is also a new fab command 'munin_node_install' to automatically
> install a munin node and configure it:
> 
> <http://knowledgeforge.net/okfn/tasks/browser/bin/fabfile.py#L433>
> 
> All nodes are now being monitored and I am therefore closing:
> 
> <http://knowledgeforge.net/okfn/tasks/ticket/321>
> 

Wow! That's great. I'm sure that will help. :-)

I should admit to adjusting (and raise possibility of having done 
something that is now incorrect) with this file on us1:

/etc/cron.d/munin

Basically, I was intending to leave munin undisturbed, but if things 
have changed I might have upset something.

What I did: commented out all the statements a couple of days ago, but 
(late-ish) yesterday uncommented them all again. I also stopped (and 
later) restarted munin-node.

Just wanted to check whether or not the former should be enabled and 
later should be running? Stats for this machine do seem to be appearing:
http://munin.okfn.org/okfn.org/us1.okfn.org.html

(Is it possible to get Apache stats showing? And is it possible to 
include PostgreSQL on that page? :-))

Best wishes,

John.

> Rufus