Installing a new package for a host in a CV and LC

I am running Katello 3.5.1 and I have CentOS 6 and 7 hosts subscribed to CentOS and EPEL repos (versions 6 and 7 respectively). On the clients when I try to install a new package that is definitely in the respective repos we get errors that the RPM cannot be found in Katello even though it is clearly there. The the most recent case we were trying to install python-docutils. I looked in Katello and found the RPM but it was saying no host applicable.

If a host is already in a ContentView and LifeCycle how can you install new RPMs? Are toy supposed to create a new CV and add the host to it?

Be sure the clients have katello-agent installed and the goferd service running. You can check that they’re getting the subscriptions for your custom CentOS repos with subscription-manager repos --list-enabled. If the subscriptions are enabled and the repos are coming from your Katello (yum repolist enabled), you should be able to find the packages in question with yum search.

To answer your question, you can add/update repositories on existing Content Views and re-publish them to provide the new content to your hosts in a given Lifecycle Environment.

If I do a add/update then republish am I not then also pulling in all the updates too possibly breaking the update LC mythology for the company?

I figured out how do what you are suggesting but I also see some major issue with it. Case in point I have clamav installed on my hosts and recently I have been having issues when trying to push out updates to the systems. The updates fail because one of the packages needed to update clamav is (supposedly) not in the repo. Which is not true.

The package is in the repo but was not originally installed on the host. Reading around I found out that if the package was not originally installed on the host katello will not install it. This is different from how RH Satellite worked.

So in my case because on of the rpms for clamav was not originally installed on the system the entire update fails. Which is not good and now forced a manual fix. Not a big deal when you only have a few machines. But when you have hundreds of hosts its really painful.