Fun fact: I procrastinated this instance so hard, that this is the second VPS I'm using for this.
The first one was cockbox I bought in may, which was very unreliable, dying too many times for a couple days each time, and I don't know why I paid for another month. A little over a week ago I broke it myself on accident. I did try to install calckey at that time (through docker) but that didn't work.
Yesterday I started over buyvm and speedran through everything while I had energy. This time I didn't use docker for calckey and did things myself. I didn't have much issues, except that I suddenly couldn't connect to my own websites through https, but today I configured snowflake and here we are now.
Also, Epik took A WHOLE MONTH to give me this domain I bought at the same time as cockbox, and the support was completely silent the entire time...
TLDR;
iska is lazy
cockbox and epik suck
network is weird
docker is not the answer.
@iska@catposter.club
If you need help you can hit me up also. I run this Misskey instance and I got basically everything worked out.
@dushman@den.raccoon.quest I basically just want to make it cool right now. I also might need to rebuild calckey since it doesn't give every user cat ears like it should. I did remove the cat mode toggle though, but that just made me give myself cat ears through postgres console.
@iska@catposter.club
Btw you did not set a CSP policy. You probably should do that. I know how to do it with nginx if you need help.
@dushman@den.raccoon.quest what's a CSP policy 🥴
@iska Content Security Policy, this shit: https://developer.mozilla.org/en-US/docs/Web/HTTP/CSP
Basically how your browser treats content coming from domains different from the one you have open.
There was a major flaw in Pleroma discovered a month (or two?🤔) ago that allowed arbitrary JS code execution by attaching a .js file to the post — there was a way to execute it for generating preview.
And everyone had to move media to a subdomain because of that 🤣
@dushman
@dcc
I think there were actually two bugs and one of them couldn't be mitigated by just adding a header on nginx side, but I didn't read too much into it — very unfun thing to do at 7AM 🤪
Well, I do in fact want to move images back and abandon the extra certificate — shouldn't be hard though, I didn't post any images on my instance since then 😅
@iska @dushman
@dcc
I think I've altered it a bit myself, I was trying to add cover art to audio files, but I don't remember if I finished it, lol
Anyway, shouldn't be to hard to rebase minor shit like that.
I was worried about DB upgrade, nice to know that there is no need for that. As it's not a VM, I can't just snapshot it and rollback if things go south, checking out old git revision and restoring DB backup shouldn't be hard either, but I've never done that, who knows what it might entail😅
@iska @dushman
@dcc@annihilation.social @m0xee@social.librem.one @iska@catposter.club
They have an RSS feed as well, it's pretty useful if you are an admin. https://pleroma.social/announcements/feed.xml
@dcc@annihilation.social @m0xee@social.librem.one @iska@catposter.club
I tend to add RSS feeds for stuff I actually run cuz they provide useful info
@m0xee@social.librem.one @dcc@annihilation.social @iska@catposter.club
You can just do a multi domain cert, you know?
@m0xee@social.librem.one @dcc@annihilation.social @iska@catposter.club
I don't think Let's Encrypt offers those for free
They do
>And everyone had to move media to a subdomain because of that
The retards way of fixing it i may add