[chef] Re: yum_package resource problem

[Adam Jacob < >]

I can't say specifically about what's happening, but I can provide some
insight into Chef. You are correct that the resources are going to be
executed in order. You don't need to have the not_if on the yum_package
resource either - that's what the package provider does internally
already, essentially.

As for /etc/munin not existing - what creates it? Is it actually part of
the post-install of the package? Is it created when the munin-node client
is started (which would explain your race?)

Either way, look at whats happening when the package itself is installed
for answers - your expectations about Chef are correct, and it's how Chef
is behaving.

Best,
Adam

On 11/29/12 2:40 AM, "jfotop" 
<
 >
 wrote:

>Hi to all,
>
>I'm facing some problems with a yum_package resource. Actually, what I'm
>trying
>to do is install a package and then configure it. The package is
>munin-node. My
>recipe contains the following resources, in the following order:
>--------
>yum_package "munin-node" do
>  not_if "rpm -qa | grep munin-node"
>end
>
>template "/etc/munin/munin-node.conf" do
>  source "munin-node.conf.erb"
>  mode 0644
>end
>------
>(chef-version: 10.14.4)
>-----
>everything worked perfectly, up until yesterday. What happened was, when
>running chef-client, the template resource through an exception and
>exited. The
>reason was that the parent directory "/etc/munin/",  which is created
>when the
>packaged is installed, was not found. When running "ls /etc/munin", the
>directory existed. And the next time chef-client ran, everything worked
>fine. 
>So, my questions are:
>- Does the yum_package resource run in the background?
>- When having 2 resources in a recipe, when does the second resource start
>executing? I had the impression that it would start execution right
>after, and
>only if, the previous resource completes successfully. What I seem to be
>getting here is the template resource starts executing before the
>yum_package
>resource has finished what it is doing.
>
>BTW: I tried using "retries" and "retry_delay" and it did work, but I
>would
>like to understand how things work.
>
>Thanks!!
>