
On Mon, Apr 22, 2019 at 01:22:17PM -0400, doctormo@...23... wrote:
Thanks Bryce (and Ian) for moving mountains to get this done.
I've been watching the ansible builds and all the issues you guys have had to deal with, and it's really great to see the hard work pay off.
The "rollout" of the new lists will be as simple as just asking people
I think the best roll out is a final shutdown message including links of where to go to sign up to the new replacement list, followed by changing the mailman2 settings in sourceforge to disallow further posting. We can start this with the board list which should be moved right away.
Actually, the process is fully scripted to do all the lists simultaneously, so it's easiest if we do them all at once. There's really only 6 lists being migrated (and 2 new lists).
People will need to re-subscribe, since SourceForge does not provide subscription lists any more.
We may have lists of email addresses used for voting.
I guess the main question here is whether we should auto-subscribe people, or if it's better to just let them sign up themselves. The downside for the latter is we'll probably end up with a smaller number of participants in the lists, but it will make the transition simpler on our end. A lot of people are pretty inactive on the lists so it may not be a bad thing to trim.
Interested in hearing other people's opinions though.
Thanks again for the amazing work!
I'm hoping that as part of the role out, we can offer up our ansible playbook to other projects in the SFC and wider open source world. Very few projects have managed to get mailman3 setups, never mind reproducible ones. Well done!
Initially I had agreed, but having gone through this process there's a lot of stuff that is particular to the host and distro version, that may make the playbook of use only as a reference. Generalizing it to suit other projects would take a bit of effort that I'm not sure has a high enough return value. Also, Recena had some concerns about making infrastructure configurations too public, for security reasons. So, I don't think publicizing this playbook is worth doing. But we could certainly share it project-to-project if there are specific people interested.
Bryce