vickleford / puppet-cloudmonitoring Goto Github PK

Hi,

within token.pp had to change

command => "raxmon-agent-tokens-list | grep ${::fqdn} | sed -n 's/.id=([0-9a-f].[0-9])./\1/p' > ${agent_token_filename}",

to

command => "raxmon-agent-tokens-list | grep ${::fqdn} | sed -n 's/.id=([0-9a-f].[0-9])./\1/p' > ${agent_token_filename}",

And change,

command => "raxmon-entities-list | grep ${::fqdn} | sed -n 's/.id=([a-zA-Z0-9]) .*/\1/p' > ${agent_id_filename}",

to

command => "raxmon-entities-list | grep ${::fqdn} | sed -n 's/.id=([a-zA-Z0-9]) .*/\1/p' > ${agent_id_filename}",

OS: RedHat 6

Appears to me that raxmon-entities-list no longer contains the fqdn in the response, even if you pass --details, which breaks when trying to store the agent id.

Since my app is not a python app nor do I have any other systems that use pip - it would be helpful to move the installation of the python-pip package into puppet-cloudmonitoring.

I'll submit a patch later

My agent_id file is empty on Ubuntu 12.04, i'll need to look into how raxmon works to figure out why this is happening.

FWIW my token is properly set.

We try and start the service before the package is installed according to these logs, will investigate further

Dec 19 16:31:54 localhost puppet-agent[18424]: (/Stage[main]/Cloudmonitoring::Service/Service[rackspace-monitoring-agent]) Could not evaluate: Could not find init script or upstart conf file for 'rackspace-monitoring-agent'
Dec 19 16:31:54 localhost puppet-agent[18424]: (/Stage[main]/Cloudmonitoring::Service/Service[rackspace-monitoring-agent]) Failed to call refresh: Could not find init script or upstart conf file for 'rackspace-monitoring-agent'
Dec 19 16:31:54 localhost puppet-agent[18424]: (/Stage[main]/Cloudmonitoring::Service/Service[rackspace-monitoring-agent]) Could not find init script or upstart conf file for 'rackspace-monitoring-agent'

I thought your module was totally whack for a while, mostly because of all of the crazy raxmon stuff going on - but I get it now, and understand why you needed to do it; although I think it sucks Rackspace haven't made forcing auto-generation of a new monitoring token possible in setup, since you can easily pass a username and api key.

In hindsight, this is a pretty decent implementation to a difficult problem (generating the token).

You should submit it to the forge so it can be installed by Librarian.

Currently works for RedHat but the rackspace-monitoring-agent has support for Debian as well.

I'm actively working on a patch for this now.