Memes about the Fediverse.
Other relevant communities:
A big thank you to all the Lemmy instance admins. We only really notice them when things go wrong. And every time something goes wrong, that means all the other admins had to perform superhuman feats so that sort of problem didn't happen to them.
But without all the other admins I wouldn't have anyone to talk to. And they test out all the updates for me.
Honest question: How large is the file size for your instance? Like, how many Gb does it all take up?
Too much.
The pgsql database is about 2 GB compressed, but it shares that with Nextcloud and I can't be bothered disentangling that right now.
But the picture storage takes the cake. I've disabled proxying images of other posts and pruned those that were proxied until then. I think. But still it takes up multiple GBs. I post many memes but not nearly enough to explain that. And because it puts all the images into a billion subdirectories it takes ages to even calculate how much storage is used. Dunno if it will be done by the time I'm done writing this post.
The picture storage is a huge black box. No idea what it's doing all the time. I can't shake the feeling that there must be better programs to handle pictures.
Performance is a big problem as well. I've only got hdds and all the constant synchronisation is hammering the database like crazy. Loading my own profile usually takes 2 or 3 tries.
Edit: It's the next morning and my pictures take up 25 GB.
You can't use like a back blaze bucket like mastodon to store all the junk?
You can, but what would that solve?
I'm not exactly sure, I just remember reading it was recommended to not fill up tons of space on your server when I was setting up mastodon, I think my server was at like 300gb last I checked.
I planned to go back and try to set it up to see what the rates were
Yeah, it solves absolutely nothing. Instead of my own server storing too many pictures I would have to pay for a second server to store too many pictures.
Know a good guide on setting up your own instance?
I think there are guides on the thing called tHe iNtErNeT
seriously though, needs reasonable hardware and no bandwidth constraints or you'll really have to watch your federation.
I used the official documentation.
I set up my instance by following the official docs: https://join-lemmy.org/docs/administration/install_docker.html
I used podman instead of docker though.
Awe...just heard about this.
Thanks for pointing to that! My main account's on slrpnk.net but I hadn't had time to look up what was going on.
Seems like [email protected] needs an internet monitoring plug. If the internet stops responding it will power cycle any devices that are plugged into it. Handy for edge devices that occasionally stroke out.
Thank you, admins!
Agreed, thank you admins!!!
👋🏻
One upside to this is that people now have a good excuse to check out Piefed!
If boost ever rolls out PieFed support, I'll be there faster than Paris breaking the warp 10 barrier
I'm still surprised you use that LW account while there are so many other Lemmy instances to choose from 😅
Lol, idk to me Lemmy is Lemmy, just with a bit of difference in the feeds depending on what instance you're on is federated with
.world is already fairly federated, too much IMO since they still haven't defederated from .ml but I digress. So I don't feel like I'm missing out on all that much.
Every time I see a hex or grad post it's always just a shit show, I just peeked over there on another account earlier and saw they were...cheering... .ee's closure soo meh
LW being too prevalent is still an issue. We see lemm.ee impact today, imagine if LW were to shutdown tomorrow.
The more people see other instances in poster's name, the more it makes them aware of other instances.
You tend to post a lot, that would be cool for you to use another instance
Sometimes home just feels like home! Having the ability to migrate a user entity independent of instance might encourage more of it though. Would be nice for comms too - eg. If somewhere like lemm.ee goes down the popular comms could be migrated to other instances and maintain continuity seamlessly to users (if such a thing were possible within activity pub). Kind of like a "forwarding address" that gets disseminated to all federated instances.
Nice! I should look more into piefed in general.
Thanks, y'all