I enjoyed the hour of the Community Hours on Friday before I had to leave.  My thanks to the team involved in the work done on this project so far, and for being extremely inclusive and open minded about the future of Uyuni.

 

My suggestions for consideration. (Two already known to the team, but adding detail and encouragement)

 

  1. Repo management improvements:
    1. Having just had to de-schedule 20 odd repos on an outgoing Spacewalk server (who’s role Uyuni is replacing) – some enhancements to the sync overview page would be welcomed. Having to click through 3 or 4 times to get to the repo scheduling page and disable them was a chore.
    2. Not being able to have an overview of when repos sync is troublesome. I like to space out syncing through a given day or days so as not to hammer bandwidth too hard. Currently I need to use a piece of paper to jot down what time each repo syncs when adjusting them.
    3. Not easily being able to see that all repos are up to date. (Obviously I get alerts when something fails, but for one-off reports and just for feeling reassured, it might be useful)

      All these may be solved by increasing the information available on the existing summary page at
      /rhn/channels/manage/Manage.do

      But add columns for:
      When last synced
      When next due to be synced / brief schedule description (Weekly day/time, Daily time, etc)
      Some toggle buttons (waves arms in a vague manner as this may need more thought) to set/move sync schedules en mass.

  1. Ansible the things
    1. I noted the comment during the discussion that Salt was not going away, so this is not an either/or suggestion. We use two technologies for automation – Ansible via AWX or CLI, and various bash/perl scripts via Spacewalk/Uyuni’s “Remote Commands”. We’ve ditched Puppet and have never used Salt. (It took some gentle persuading to get Uyuni adopted because of adding another stack). We don’t intend to use Salt for anything beyond what Uyuni needs under the hood. (And to be fair, salt-minion has already proved to be better behaved than rhnsd and osad so it’s a good choice)

      My suggestion is adding the ability to run ansible playbooks on remote machines through Uyuni, both ad-hoc and in schedules. Perhaps copying the “Remote Commands” tab and being able to use either/both:

                                                               i.      Cut and pasted playbooks

                                                             ii.      Select playbooks from a local or remote repository.

  1. Centos as a Uyuni host.
    1. Again, it took some gentle persuading to get a Suse box on our network as we’re a Centos shop and we know Centos. However, I know Stefan is working on this already and there’s little to add to the words said by Neal and others, other than to wave encouragingly from the sidelines.

 

 

Simon