Update: I've just seen this:
I run lemmy.world on a VPS at Hetzner. They are cheap and good. Storage: I now (after 11 days) have 2GB of images and 2GB of database. (https://lemmy.world/comment/65982)
A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don't control.
Beginning of January 1st 2024 this rule WILL be enforced. Posts that are not tagged will be warned and if not fixed within 24h then removed!
Update: I've just seen this:
I run lemmy.world on a VPS at Hetzner. They are cheap and good. Storage: I now (after 11 days) have 2GB of images and 2GB of database. (https://lemmy.world/comment/65982)
Our smaller instance that has been federating for a bit more than a year now (started in March 2022) is now at 2.4gb for the database and 7gb for the image storage (which probably needs some clean-up from previous image spam waves).
@poVoq @ndr this will surely grow a lot, if you look at other activity pub compatible systems you'll see a huge grow, it depends on the retention of "old" post and media, if you say just store all for a year you might keep it smaller, but if you want to replace #Reddit or so it would be better to keep stuff a bit longer, but then on the other hand the #Fediverse is probable not meant to store stuff for long term.
On my Friendica node I have a rather short period to store foreign posts and media, and my storage is only about 47 GB, most of the media is stored in the database as well (easier and faster to backup, much slower to retrieve) and it is a single user instance with just a handful of bots besides the account I write from.
Im gonna wait before I host a public instance. This is the first experience with an actual bigger user base for most instances.
I would suggest that people don't just willy nilly start an instance without long term planing for database growth and media storage.
Oh yeah, I'm still learning and planning when/if I do it.
OTOH, we shouldn't scare people off doing it either LOL
You are right. But giving users a bad experience is also not good for future reputation. Users are fast with giving something they had a bad experience with, a negative stigma.
Do you plan on allowing other users? If so every image they upload anywhere will be hosted from your instance. You would need a long term plan for continued storage if you do.
If it’s only yourself then not much. You would need space for your uploaded images and the database. Worst case you have to purge communities to free up DB space and re add them. It only tracks communities from the moment you add them. It doesn’t pull the entire history, and associated db size, into your instance.
I can pull the real numbers when I’m not on my phone but my database and maybe 150mb now. You can see all the communities I follow here so get a relative idea: https://lemmy.cablepick.net/communities My instance has been up for 4 days. I should start tracking db size growth to give others an idea of what to expect.
If I follow through with this, it won't be very soon. I'm just gathering information to see whether it's worth it.
Anyway, I was thinking of keeping it only for me and a few more users, maybe (definitely not open for everyone).
Can one set up their instance to delete posts amd associated media after x number of days?
You would have to run raw database commands to clear followed communities’ posts. It would be possible with some experimenting. The only local media associated with non-local communities is generated thumbnails so those would have to identified and purged from pict-rs.
One thing to note is that pictrs supports hosting your images on in object storage (e.g. AWS S3 or Backblaze B2). That can save a ton of money on storage costs. 1TB of images on B2 only costs $5/month.
EDIT:
looked at my instance last night that I'm only subbing to other communities. It's been running 2.5 days at that point. My VPS with Ubuntu 22.04 is at 5gb total. Next time I ssh llI look at the database size. I think I can confirm pictures are still loading from source instance (lemmy.ml went down for a brief second last night and all my images with them stopped loading).
it's my understanding that you won't make a full copy of everything. you'll only copy the communities that are added and you won't be copying the full history.
I'm unsure how far back it goes but it might be just the one day. i saw talk of changing it to go back as far as someone actually scrolled and pressed next but that was just an idea at this point.
Also it doesn't copy the images as Mastodon does, only the text objects.
oh right, that's a big one there. that images stay on the origin instance and they linked from there
I actually kind of like how Akkoma does this. It creates a constant size proxy with nginx, and all the images come from predetermined host instead of all over the net. It's a good mixture of not using tens of gigabytes of space and still spreading the load a bit between the instances.
For this reason, I assume it should be feasible to self-host an instance that sync with most communities.
For reference, Wikipedia (en) takes up just over 20 GB (https://en.wikipedia.org/wiki/Wikipedia:Size_of_Wikipedia), so I assume much less than that for Lemmy (with the current userbase).
Okay, I see, thanks.
But for example, do we know how much data lemmy.ml stores?