I've been thinking a lot about running mcollective on my production
servers recently, especially about its nice integration with puppet
(using facts and so on). However I've got some concern about its
scalability - how publish/subscribe middleware scales in terms of
speed and flooding the network.
configure puppetmaster for this), but this AFAIK
requires you to maintain multiple instances of your modules and
classes.
-Regards
On May 3, 9:26 am, Patrick wrote:
> On May 2, 2010, at 9:38 PM, linuxdatacenter wrote:
>
>
>
> > Hi,
>
> > Just want to know how you use puppe
Hi,
Just want to know how you use puppet to cover a scenario where you
have a production and backup datacenter. Both environments should be
similar but not identical (just a dumb example - your ntp setup is
probably different in each). I think it's desirable to have just one
implementation of modu
Hi,
I've written a small tutorial on how to deploy images for puppet using
systemimager. Please check it out:
http://linuxdatacenter.blogspot.com/2010/05/systemimager-puppet.html
-Linuxdatacenter
--
You received this message because you are subscribed to the Google Groups
"Pu
Hi,
Just a survey how you guys model different aspects of your systems. I
mean: in my world every node is determined by 3 factors:
-> role in the datacenter (like www-server, workernode, db-server,
etc.)
-> website membership (I run several different websites with different
webserver configs, db,
