!news@lemdit.com
Stay up to date with the latest happenings.
!news
@lemdit.comhexbear.net is "A leftist social platform centered around community building through discussion, shitposting memes, and sharing content."
They are fairly large:
hexbear.net were not a federated instance until recently, when they announced their intent to start federating with a select group of other Lemmy servers.
lemmy.world claimed to see ulterior motives behind this intention and decided to pre-emptively block hexbear.net from being able to federate with them.
It's becoming increasingly apparent that the largest Lemmy instance thinks it has a duty to protect its users against anything that might cause upset, sometimes before it even happens.
lemmy.one has been down for a while now. I've personally noticed that it hasn't been working for over 24 hours, some comments suggest it may be longer than that:
lemmy.one is one of the larger Lemmy instances, with around 7.3k users.
The site does load, however it displays a Lemmy error. This shows that the Lemmy UI is still there and working, but the Lemmy back-end server process is not.
Based on this, it doesn't look similar to the disappearance of vlemmy.net and it's most likely just a technical issue.
My guess is there are 3 likely scenarios for what went wrong:
This is a very long time to be struggling with a technical issue and the lack of communication is concerning.
It could be that:
I expect they have regular DB back-ups like everyone else and would've rolled back to a previous snapshot by now. I think the most likely scenario is that the admin just isn't available to deal with it.
To fix this requires actual server access, so even though lemmy.one had several Lemmy admins, it's unlikely that all of them have the required access to deal with this issue.
It would be interesting to hear if anyone managed to get in touch with lemmy.one's admin.
Edit: It's back: https://lemdit.com/post/294679
This version brings major optimizations to the database queries, which significantly reduces CPU usage. There is also a change to the way federation activities are stored, which reduces database size by around 80%. Special thanks to @phiresky for their work on DB optimizations.
The federation code now includes a check for dead instances which is used when sending activities. This helps to reduce the amount of outgoing POST requests, and also reduce server load.
In terms of security, Lemmy now performs HTML sanitization on all messages which are submitted through the API or received via federation. Together with the tightened content-security-policy from 0.18.2, cross-site scripting attacks are now much more difficult.
Other than that, there are numerous bug fixes and minor enhancements.
Full release notes can be found here: https://github.com/LemmyNet/lemmy/blob/main/RELEASES.md
I've noticed that traffic from join-lemmy.org dropped off entirely for the past couple of days - it was patchy for a while before that.
It looks like our instance has disappeared from the list, along with others such as lemmy.nz
I'm guessing there are probably more servers missing, but I don't know about them to be able to confirm.
What I do know is this has nothing to do with our uptime and I can see the join-lemmy.org crawler touching base successfully every hour.
I've submitted a bug report for it.
join-lemmy.org is an important part of the Lemmy ecosystem so I do hope this gets identified and resolved.
In the meantime we're a very well kept secret.
lemmy.blahaj.zone has defederated from lemmynsfw.com after discovering a community which they believed had CSAM (Child Sexual Abuse Material).
They demanded that the community and its members be purged, which lemmynsfw.com refused stating that they couldn't find any evidence of CSAM.
Both sides think they're in the right on this one and I expect more instances will want to react one way or the other.
It has been another eventful week.
What does this have to do with Lemmy, you ask? Plenty:
As of writing this, lemmy.fmhy.ml is down and it's unclear what will happen to lemmy.ml and lemmygrad.ml.
lemmy.fmhy.ml have announced that user data is safe, though changing domain names is a huge pain for re-federation and nobody really knows how that will work yet.
If you are unable to access communities that used to reside on lemmy.fmhy.ml - this is why.
In general, it's a good idea to stay away from country specific TLDs when hosting an instance, epecially if you don't reside in those countries.
This one unfortunately comes with the territory once you reach a certain size. If you've been struggling to reach lemmy.world recently, this is probably why.
Fixing this kind of issue will lead to massive performance benefits for large instances and improve resilience.
Related:
What did I miss?