As I was browsing lemmy and the fediverse at large, this question kept popping into my head.
Since multimedia files have a much bigger footprint than raw text, it made me feel worried since as time goes, massive resources will be needed to keep up with the big data coming in.
I do wonder if the instances have taken the route of the cloud and just decided to put all of it in something like AWS S3? Or maybe they use self hosted storage with something like minio for object storage?
This will differ greatly from instance to instance. The people running lemmy.world have published some info on their infrastructure. My instance is running on a rather small VPS with 100GB storage, but I will have to rethink my solution rather soon as images and videos from my subbed communities [Edit: which are stored on outside sites] are eating around a gigabyte per day and I think this is likely to increase.
Edit: I want to clarify that I was partially wrong - Lemmy only locally caches content which is hosted on outside sites (e.g. imgur). It does not cache content that was directly uploaded to another Lemmy instance and just embeds the source media.
Is it caching the entire image from the post or just the thumbnail?
Everything. It does some re-encoding when it retrieves content from other instances and you can set limits for pictrs (the software Lemmy uses to host media) regarding file sizes etc.
Edit: I was partially wrong about what is cached, see my original comment
When I was looking into hosting my own instance I thought I saw an option to disable media file replication entirely so that they would always have to be fetched from their home instance.
That would be great to know, any chance you remember where you read that?
It’s possible that I’ve misunderstood. And it’s also important to note that I was looking into this for the purposes of creating my own, single user instance. I wasn’t planning on posting to my own instance, just using it as a single logon where I could control what other instances I federated with.
Here it mentions not installing pict-rs and removing its configuration if you don’t need image hosting. My interpretation at the time was that it would mean that no images would be hosted locally on my instance. But that was very early on before I understood more about federation, and now I realize that it may in fact also mean that any content coming from federated instances could have images broken, not that it would load the images from the remote instance. So now, I no longer think that this is a solution for not syncing images, but I’m not at all sure of that.
No, but I bet I could find it again if I hadn’t just imagined it and made it up for this comment. Give me a few.
Thank you for your work for the community! I think with more people using lemmy, we should also as users lookout for the infra we are using because the admins are not a mega corporation ready to spin up infinite resources.
No need to thank me, currently I am the only non-bot-user of my instance and do not allow registrations 😅
Many of the bigger instances have links to donate to their operators, but I am doubtful that relying solely on donations will be enough in the long run.
Since you’re the only one, you might consider setting an expiration on the media so your local storage serves as more of a cache. Like, I’m sure you’re far more likely to revisit a recent thread than a super old one, and as long as the original instance is still around you could redownload the media. This might require software patches though idk
Maybe have users use an outside image provider, like imgchest or gfycat or whatever?
Edit: I am partially wrong. (See below)
They’re stored on their host Instance. Only text is copied across instances.
That is not true. As long as a user on your instance is subscribed to a community, the media content of posts [Edit: only posts linking to outside sources, e.g imgur] of that community is stored locally on your instance as well.
This, of course, only applies to media which is uploaded to Lemmy, links to media hosted externally are not downloaded.
See this issue for more context.
Edit: I want to clarify that I was partially wrong - Lemmy only locally caches content which is hosted on outside sites. It does (should?) not cache content that was directly uploaded to a Lemmy instance and just embeds the source media.
I think this could be a ticking DOS time bomb.
Someone manages to spam upload massive files to the largest Lemmy instances could wipe out a ton of smaller ones.
Not to mention scalability wise this seems like a nightmare… eventually the largest Lemmy instances will have petabytes of media data with 100s of gbs coming in per day, giving other instances no chance to sync with them.
I think the system architecture needs a significant review. This won’t scale.
I agree. It’s also a tremendous waste of resources. I’m all for redundancy (like CDNs), but this seems incredibly poorly thought out. If Lemmy (as a whole) every scales to the size of other social media, the space requirements will start to become unreasonable.
Why wouldn’t something like symlinks be implemented? Not saying specifically use symlinks, but there has to be a similar, better way.
The obvious way would be to just not cache content locally and always link to the source instance. While this would concentrate the strain immensely, it would also greatly decrease the storage space used by all other instances.
There might also be other viable alternatives such as using a CDN and having it selectively cache content which is requested often etc.
~~As of now, Lemmy does not support either, though. ~~
Edit: I want to clarify that I was partially wrong - Lemmy only locally caches content which is hosted on outside sites. It does (should?) not cache content that was directly uploaded to a Lemmy instance and just embeds the source media.
