Subscription checks very complicated about ansible-hana-examples HOT 4 CLOSED

@tbludau There is some good work that can be done here.. Both in specific tasks and structure.

Many of these tasks where the shell or command modules are being used for subscriptions can be remediated using either the redhat_subscription or yum_repository modules.

As far as structure we should be able to reduce some of these plays by enforcing a desired state around subscriptions rather than doing checking like so:

- name: Reset all previous repositories
  shell: |
    subscription-manager repos --disable="*"
    num=0
    for r in /etc/yum.repos.d/*.repo; do
      num=$(( num + $(grep  -c "^enabled.*=.*1$" $r) ))
      sed -i 's/^enabled.*=.*1/enabled = 0/g' $r
    done
    echo "$num"
  register: repo_cleanup
  changed_when: repo_cleanup.stdout > 0
  failed_when: repo_cleanup.rc > 0
  when: force_repo_reset

We shouldn't need to do loops in tasks for each repo, the yum_repository module should be able to disable all repos. We should also try to get to a state with this role assuming that we either want to force a repo reset every time or not before this task.

If there are instances where the answer is 'it depends' we should try to be more selective in resolving repo conflicts

Lastly, I think we can restructure the suggest roles into either 'self-configured' or 'subscription-based'. We should be able to perform a couple of tasks based on wether a Sat 6 URL var is present or not, then proceed as normal.

Thoughts?

from ansible-hana-examples.

@tbludau I think we can close this task -- there is a lot of cleanup to be done in issue #13 so I can resolve some of these.

from ansible-hana-examples.

@tbludau question.. are we setting a strict level on ansible version? There's currently a new rhsm_repository module in the development stream for ansible 2.5 that might clean up some of this too.

from ansible-hana-examples.

hey @kyleabenson thanks for your input and sorry for my delayed response. as you descibed we can close this case, because it should not be part of the ansible deployment and brings a lot of erros in aws and azure environments.

from ansible-hana-examples.