A general update for those interested.

First, a quick announcement that for the next 12-24 hours our wiki will be unavailable while the domain record is updated. I trust that this won’t bring you too much despair.

A lot has been going on behind the scenes over the past couple of weeks. I figured our faithful users might be interested in the who/what/why, so here’s my best effort at explaining the challenges we’ve had up until now.

Going back to the beginning, when I first became aware of Lemmy, @merikus was in charge. He did a great job of setting the instance up, putting rules in place, and recruiting members. The hosting for our instance was done by a company called K&T Hosting, and we were a beta instance, which meant it was free.

Fast forward a bit, and Merikus needed help running the place. Even with K&T Hosting running things, there was still a lot of little maintenance to be done here and there. I dabble in computer stuff, so I volunteered.

A good auto-post bot didn’t exist, and the one that Merikus found in the early days broke after one of the Lemmy updates. That coincided with the Lather Games on Reddit during which Merikus, Djundjila, and myself were very busy, and @PorkButtsNTaters666 did an excellent job of keeping the thrill alive by posting daily threads. Thanks again!

As I became more involved, I started to think it’d be nice if we had more control over the server. I found a different auto-post bot which “worked”, but that was hosted on my little home server, and not a good long term solution. Now we have a nice new bot which is updated quite often and has a very responsive developer.

So the desire to host all of this on our own rented server using Docker has existed for 5 months. Fast forward to the past couple of weeks, and our hosting service started having problems. Every instance hosted by them experienced the same things.

K&T was very active in getting things back to normal, and they have a good plan in place to prevent the same problems in the future.

We like K&T Hosting (hello if you’re reading this!), and I’ve been impressed with the customer service. The hickups of the past week or two are NOT why we decided to change hosts, but it was a catalyst for @djundjila and I to finally get to work doing it. We simply wanted more control, and saw some benefit to hosting it ourselves.

Now our instance is hosted on a Hetzner server. We have better control over the data and updates, but we aren’t immune to problems. For example, yesterday I thought it’d be fun to change the icon for the instance. I uploaded a picture that I stole from @djundjila, and BOOM, Lemmy crashed.

The part of Lemmy that handles pictures did not like whatever I uploaded. I ended up having to manually modify the database, and the instance was down for about 45 minutes while I scrambled to figure it out.

So anyway, I’d like to mention again that K&T Hosting was good to us, did an excellent job hosting, and we wouldn’t hesitate to recommend them. Our instance suffered an occasional problem while being hosted there, just like it’ll have the occasional problem on it’s new hosting platform. It’s just personal preference for the Admins to have it hosted somewhere else.

I learned a ton from @djundjila while migrating the data. From dumping and restoring databases, transferring files between servers, organizing multiple domain names, and making sure everything is secure… I couldn’t have done it without his help.

A general overview of how things are managed now, for those curious:

1. Lemmy and associated software is run in Docker, installed on a server running Debian.
2. Portainer is used as a web interface to manager the docker settings, updates, etc.
3. Nginx Proxy Manager handles the security and SSL certificates.
4. The wiki is also hosted on the new server, instead of at my house.
5. The auto-posting bot, called LemmySchedule, is also hosted on the new server instead of at my house.

Thanks for reading,

• walden