Community sites migration

Community Ops
#1

I’m opening this thread up before the meeting so we can get the ball rolling. Terraform is complete (just needs a review from someone), Ansible is pretty much ready to go (we just need to sort out a derp tanner made when moving repos).

Me and Tanner spent some time yesterday documenting what software is running on OVH shared so we can get started on dockerizing the most common services. We’re hoping to do something similar with the MediaTemple server today.

Is there anything else we need to consider before we start moving people over? @majken, how much notice should we be giving to communities?

(note this is internal because I’d rather not show the world which domains are running outdated, exploitable software)

#2

What are we giving them notice to do? ie what should they be preparing for? Or is it just so they know that their site might be down for a bit?

My answer hinges on that, though I would say generally a week given the pace communities move. That is time for them to post a notice that their site might be down and for people to see it.

#3

For WordPress sites we might disable plugins that interfere with MainWP (backup plugins etc), unmaintained or old versions of plugins.

For other software they will have to make sure their instance is running the latest version of that software otherwise we can’t guarantee it’ll work properly after the migration.

Some websites were running custom code which wasn’t in a Git repo, so we may need to reach out to them to figure out how it works so we can help them dockerize/dockerize for them.

#4

How much of this falls under stuff you need to do, and how much is stuff we need to do?

I’m guessing for WordPress sites, once there’s a new instance spinning, we work with the community to migrate the content, and you flick the DNS switch when it’s ready?

For custom code base stuff, it might make more sense for you to work directly with the community.

If starting with WordPress sites will be easy, I’d love to do that.

#5

The stuff we need to do overlaps with the stuff you need to do, otherwise it’ll take much longer to migrate. What you can do is finish off documenting by letting us know what software is running on the MT and other servers.

#7

@tanner is starting to spin up the production MainWP and some wordpress instances and we’ve detailed the immediate next steps here.

@tad @majken we’re going to be opening up technical bugs for each migration and I think it’ll make sense to include the link to them when you contact the communities.

e: I’ve also removed the pad in the OP and moved the thread out of internal

#8

We have filed bugs for all of the sites on OVH Shared. @tad @majken, when you contact the communities, could you also include a link to their bug to verify if we have found all of their services?

https://bugzilla.mozilla.org/showdependencytree.cgi?id=1267162