17:37:03 <raub> #startmeeting
17:37:03 <Tribot> Meeting started Mon Apr 12 17:37:03 2021 UTC.  The chair is raub. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:37:03 <Tribot> Useful Commands: #action #agreed #help #info #idea #link #topic.
17:37:29 <raub> #chair bdmc jeoff
17:37:29 <Tribot> Current chairs: bdmc jeoff raub
17:37:46 <raub> #topic 1. Check for Agenda Additions
17:38:03 <bdmc> Nay
17:38:16 <raub> #topic 2. Current Topics
17:38:46 <raub> For this month's meeting, I might have been able to record but will see if it worked only later int he week
17:39:14 <raub> Next month is elections. I will send out an email about that to the list
17:39:17 <bdmc> I look forward to it.
17:39:55 <raub> Let's see how it goes
17:40:04 <raub> Also, regarding Pilot OS Upgrade,
17:40:20 <raub> Did anyone check the upgrade plan on the wiki?
17:40:39 <raub> https://steering.trilug.org/wiki/index.php/PilotUpgradePlan
17:40:40 <bdmc> I had trouble finding it, since it is not "publically" visible.
17:41:42 <bdmc> ( or not in the main menu )
17:42:02 <raub> I have no idea where it needs to go
17:42:23 <raub> I first found it because it was on https://steering.trilug.org/wiki/index.php/PhysicalServerLayout
17:42:25 <bdmc> I don't really, either.  Perhaps "System Administration?"
17:42:44 <bdmc> I thought that I looked there, and didn't see it.
17:42:49 <jeoff> Just took a look. We split pilot into core and user services, then follow your MOP. That's good enough to create playbooks around I think
17:42:51 <raub> Bottom
17:43:02 <jeoff> If we can assign IPs to these VMs, the automation can get easier
17:43:44 <raub> It depends on which IPs. We can control the internal IPs
17:44:18 <bdmc> Yes, my expectation was that we would assign public IPv4 addresses to the two replacements for Pilot, and Moya.
17:44:57 <bdmc> At the moment, Moya and Pilot are trying to share the same one.
17:45:32 <jeoff> Internal IPs are fine as long as we can SSH to it somehow. Just need something for scaffolding
17:45:43 <jeoff> Probably worth having a directory of services too
17:45:56 <bdmc> Agreed
17:46:14 <bdmc> We have taken stabs at that several times.
17:46:18 <raub> jeoff: there is a space in https://steering.trilug.org/wiki/index.php/PilotUpgradePlan
17:46:23 <raub> for that
17:46:47 <raub> https://steering.trilug.org/wiki/index.php/PilotUpgradePlan#Configuration_and_Services
17:47:13 <jeoff> Ooo nice
17:47:24 <jeoff> And you called out Ansible too
17:48:11 <raub> It all depends on how we want to push it from where
17:48:56 <bdmc> Do you want to make "Ansible" the decision?
17:49:05 <bdmc> Any other candidates?
17:49:13 <raub> Don't care
17:49:25 <bdmc> That's not a decision.
17:50:14 <raub> There are 3 people here in this talk. If you prefer something else, make an argument. I am not married to any solution
17:50:27 <bdmc> jeoff?
17:50:45 <jeoff> Ansible
17:51:46 <bdmc> Done
17:51:56 <raub> Moving on
17:52:42 <bdmc> Something that Jeoff said in passing.  Should the Ansible "master" be somewhere off site?
17:53:22 <bdmc> ( Sorry, that was raub. )
17:53:47 <raub> IMHO all that matters is to keep the private stuff private. The rest could even be in github
17:53:57 <jeoff> I've heard good things about Ansible Tower/AWX, but that would require complex configuration. Maybe we have a jump box within the network where we run playbooks
17:54:59 <bdmc> Moya?  Something completely different?  ( another new machine? )
17:55:06 <jeoff> another new machine
17:55:18 <bdmc> Single-purpose?
17:55:31 <bdmc> ( tiny machine, in other words )
17:55:48 <jeoff> yeah
17:55:56 <jeoff> 1c/2Gi would probably be fine
17:56:06 <bdmc> Exactly
17:57:24 <raub> I suggest most of the playbook can be pubicly available and the private stuff can be local
17:58:43 <jeoff> raub: we're running out of time, but I'm curious what parts you think should be public
18:01:33 <bdmc> Who wants to create that machine?  I could do it, and could leave the Ansible part to either or both of you.
18:01:39 <raub> jeoff: we can worry about making stuff public later, but installing packages can be public. And templates, which are then fed local vars that are not exported to the repo, can also be public
18:01:55 <raub> Only reason is so others can use it, including for a future class
18:02:17 <raub> And it forces people not to be lazy and have secrets hardcoded
18:04:15 <jeoff> If we use Github, the future class would have access once they're invited to the TriLUG Org and we can leverage a lot of cool features
18:04:57 <jeoff> But that's something that's probably outside of the scope of this. I know we need a new git repo (where it's hosted doesn't matter for now) for this body of work
18:05:11 <jeoff> bdmc: If you build it, they will automate
18:05:17 <bdmc> B-)
18:05:24 <bdmc> Will do.
18:05:47 <raub> jeoff: exactly
18:06:27 <bdmc> Theoretically I am only employed at the moment in the "job hunting" job, so I should be able to get this done in the next day or so.
18:07:19 <raub> bdmc: If you need help, let me know
18:07:23 <jeoff> bdmc: once that's done, feel free to drop a note to the Steering List so I can start my piece
18:07:43 <bdmc> That was what I was planning.
18:09:11 <raub> If this is under control, I would like to move to the next topic
18:09:30 <jeoff> I have a relatively free week, so you should see some commits to the internal git server throughout the week. Planning on having something presentable on Friday. Can someone make the empty git repo?
18:09:52 <raub> As in github?
18:10:12 <jeoff> nope, the same way we do it for Jekyll
18:10:33 <jeoff> don't want to bite off the Github part until we can plan what it looks like a bit more
18:10:43 <raub> Oh. I think I can but  will reach out to Allan first
18:11:30 <raub> jeoff: just be extra careful to use, say, something in host_vars so only variables are used in the reoles
18:12:15 <bdmc> raub: That sounds too much like planning ahead! B-)
18:12:46 <raub> What a concept!
18:12:47 <jeoff> raub: I'll be careful
18:13:53 <raub> Excellent
18:14:10 <jeoff> I look forward to sharing on Friday
18:14:37 <raub> jeoff: sure thing. As soon as I have the repo I will let you know. I also want to see what is there first
18:15:11 <jeoff> How much of a lead time do you want?
18:15:39 <jeoff> raub: also, can you make the primary branch `main` instead of `master`? There's a big push for that and I'd rather not have to fix that later
18:17:50 <raub> I do not know if that is supported by the setup
18:18:14 <raub> Hence my github suggestion. They do offer private repos
18:19:39 <jeoff> So should I invest in the Github setup?
18:20:48 <jeoff> Should be free https://github.com/pricing#compare-features
18:23:33 <raub> IMHO yes
18:24:36 <jeoff> Alright, I guess that's more stuff to run by Alan
18:25:01 <raub> Yep
18:25:10 <raub> ?
18:26:15 <jeoff> Why the question mark? I figured we should get his thoughts on it before we did it, because he'd probably be the Org admin
18:26:19 <jeoff> just like the other accounts
18:28:05 <raub> K
18:28:58 <jeoff> anything else?
18:29:03 <raub> I am good
18:29:56 <jeoff> Cool, this is probably a good place to end
18:31:17 <raub> Exactly
18:31:23 <raub> Alright then
18:31:27 <raub> #endmeeting