vagrant up

Let's see what happens.

😕 Hrm... I just made it past ansible-galaxy unscathed and didn't have to force provisioning. I'll let you know when it's totally done if it's fine.

Did you delete your roles/external directory?

If not then you didn't have to re-download all of the geerlingguy roles and you only downloaded our 6 new roles. Hence not hitting the 21 limit.

Ah, ok. Let me try that.

🤷‍♂ I dunno, just tried it twice, clobbering roles/external and it worked fine both times.

What dark pact did you make?!?!!?!

Me and Beezlebot go way back.

But it's probably just my ultra-slow internet narrowly avoiding whatever rate limit they've set.

Please ignore my previous comment. vagrant up works with this PR. No errors. Basic CRUD (create a resource, sync to Fedora, Triplestore) works as expected.

Are we cool to move this in, then?

@dannylamb Yes, Works as expected for me. In addition to this one, if 40 and 44 can be merged, then claw-vagrant can be decommissioned.

@whikloj @Natkeeran @seth-shaw-unlv Ok, before this goes in then, let's take a look at what we've got going and see if there's other stuff we can sneak in first before I unleash some git-fu on all the repositories for the split out roles.

Wait?! What git-fu are your doing?

The plan (as I understood it) was to NOT do anything to claw-playbook until we got this merged and the external roles in use. Then I was going to start on PRs to clean up all the variable overlap that was happening in the various roles.

So any PRs for the individual roles could be recreated on their specific repository (ie #44), but then it would be easier to test as we would actually be using the external roles here.

@whikloj Yes, that approach seems good to me.

I was totally going to just force push whatever changes have already rolled through to the individual repos and merge the PR after testing again. Just sense it's stuff we've already reviewed. Subsequent work would still go through the normal workflow for the applicable role(s).

@dannylamb But you shouldn't need to force push anything. I rebuilt all the external role repositories as exact copies of what is currently in claw-playbook (I had to add the meta/main.yml to some). But otherwise this PR should merge without any issues and once that happens we are moved to using external roles.

Just going to throw this here to tie the two discussions together. I can't get the roles to come in cleanly, same error as Jared had. I made a PR on his PR here: whikloj#1 that pulls them from git instead of galaxy and that works okay for me. I dunno if thats an okay solution for everyone.

So this comes up fine for me after two things:

1. switching to git for some roles: whikloj#1
2. commenting out FITS: Islandora-CLAW/CLAW#747
   • PR for this here: #46

I guess I can't merge this anymore because I'm implicated in it now, but I'm satisfied and would merge this after the two PR above are merged.

I'm going to have to rebase this PR after the FITS removal, so do we have any issues with switching the galaxy roles to Github repos?

If we don't I can merge @jonathangreen's PR.

Personally I'm a little concerned about the fact that Galaxy seems to be our failure point.

@dannylamb I just went ahead and merged @jonathangreen's PR. It solves the problem for him and me (maybe a OSX issue) and now we are good to go.

@dannylamb Can we merge this one!