ordnung, to random German
@ordnung@chaos.social avatar

There will be a short downtime for maintenance in a few minutes.

michael, to random
@michael@thms.uk avatar

PSA for admins for glitch-soc:

The latest version (merged in today) requires you to set up three environment variables before you can start your server.

Details can be found here:

https://github.com/glitch-soc/mastodon/issues/2525

Jerry, (edited ) to random
@Jerry@hear-me.social avatar

#MastoAdmin #MastoDev

UPDATE on 216.106.86.232 :

Every few seconds, it's doing the following GETS. I'm getting about 2,000 calls per hour from this IP.

https://hear-me.social/packs/media/images/friends-cropped-40409bf96999902fa49757ebce824ce6.png

https://hear-me.social/api/v1/streaming?stream=public:local&access_token=O6luKbPRJV0FlTvhCSqn6z4wNwkUJMxUPgD7xMn4fsc

https://hear-me.social/api/v1/announcements

https://hear-me.social/api/v1/notifications?exclude_types[]=follow_request&since_id=79462

Is anyone else getting traffic from 216.106.86.232 ?

I will open a report with the ISP

=======================

I'm putting this under #FediBlock because I've noticed that since 12 PM EDT yesterday (30-Apr-2024), API requests coming in from 216.106.86.232 have been returning 500 errors constantly from my server. No other 500 errors were being returned.

I blocked the IP at the Cloudflare firewall and now the 500 errors have stopped. There were more than 93K 500 return codes in total until now!

I have no idea what the requests are, and it's too late at night to do more research, but I'm passing this along to other Admins as a warning and to see if anyone else is seeing similar behavior.

@Gargron

Here's what I see in Cloudflare for the last one hour before blocking the IP address. Notice all the requests returning 500 are from this single IP. There are no page views. It's all API requests. And 500 errors stopped after I blocked it.

Any insights would be appreciated.

admin, (edited ) to random
@admin@hear-me.social avatar

UPDATE on 216.106.86.232 :

Every few seconds, it's doing the following GETS. I'm getting about 2,000 calls per hour from this IP.

https://hear-me.social/packs/media/images/friends-cropped-40409bf96999902fa49757ebce824ce6.png

https://hear-me.social/api/v1/streaming?stream=public:local&access_token=O6luKbPRJV0FlTvhCSqn6z4wNwkUJMxUPgD7xMn4fsc

https://hear-me.social/api/v1/announcements

https://hear-me.social/api/v1/notifications?exclude_types[]=follow_request&since_id=79462

Is anyone else getting traffic from 216.106.86.232 ?

I will open a report with the ISP

=======================

I'm putting this under because I've noticed that since 12 PM EDT yesterday (30-Apr-2024), API requests coming in from 216.106.86.232 have been returning 500 errors constantly from my server. No other 500 errors were being returned.

I blocked the IP at the Cloudflare firewall and now the 500 errors have stopped. There were more than 93K 500 return codes in total until now!

I have no idea what the requests are, and it's too late at night to do more research, but I'm passing this along to other Admins as a warning and to see if anyone else is seeing similar behavior.

@Gargron

Here's what I see in Cloudflare for the last one hour before blocking the IP address. Notice all the requests returning 500 are from this single IP. There are no page views. It's all API requests. And 500 errors stopped after I blocked it.

Any insights would be appreciated.

paul, to random
@paul@oldfriends.live avatar

Like clockwork, on Wednesdays, my Sidekiq pull Queues jump to 3K most of the day. After a half a day or so, it comes back around.

I wonder if there is a post dump by a bot or some task my server is doing? I am thinking a weekly task must be running.

I need to dig into my logs, crons, etc.

.

rolle, to mastodon
@rolle@mementomori.social avatar

:skull360: We have now successfully upgraded to v4.3.0-alpha.3+mementomods-2024-05-01. The server has Mastodon Bird UI 2.0.0rc37 installed with no further changes at this time in the UI.

The branch is now synced and up-to-date and 158 commits ahead from the last version 2024-04-06.

I love running the newest version so I can test ahead in production for the upcoming Mastodon Bird UI versions. I do have my local instance with the production db, but it is far more fun this way and it feels good to have new things.

This "testing in live" is also because it's the perfect way to make sure that I can help contributing to Mastodon and spot bugs before they end up in the final rc, beta or stable versions. There most probably will be some minor bugs, but the earlier we tackle them the better.

As usual, please tell me about any strange things you may encounter with this Mastodon version. We're running bleeding edge after all.

I hope you enjoy your time here. Feel free to send me a message if you have anything in mind. :bunhdheart:

gunchleoc, to random
@gunchleoc@mastodon.scot avatar

My Elasticsearch is killed sometimes because it uses too much memory. This is what I did to reduce the memory it uses: https://co-shaoghal.net/index.php?page=173&offset=0#section246

I also have a script in place to restart any failed services: https://co-shaoghal.net/index.php?page=173&offset=0#section240

simounet, to random
@simounet@mastodon.simounet.net avatar
paul,
@paul@oldfriends.live avatar

@simounet @ricard

My only issue with the removal tasks is, if you use Elasticsearch and run 'tootctl statuses remove', does it remove statuses from the search database?

I don't run that command because I want to find posts sometimes. I'm using my instance as my primary Google Search replacement. I am getting more useful finds here than Google these days.

I've asked before to no avail. I'd really like to know if it does remove stuff from Elasticsearch.

Whiskeyomega, to random
@Whiskeyomega@cupoftea.social avatar

Well this may take some time

ian, to random
@ian@phpc.social avatar

PSA: Vultr cloud storage costs are now down to $6/TB/mo.

Backblaze B2 still has the edge on data transfer costs (they bundle data transfer up to 3x the storage amount) but IIRC bandwidth is pooled with cloud servers on Vultr anyway, so you probably won't pay for outbound anyway, and it's one less vendor in the stack to manage. Which is nice when you have better things to do than babysit various vendor accounts as part of running an instance :)

vga256, to random
@vga256@dialup.cafe avatar

😭 for nearly a year i've had nothing but problems uploading videos to my instance. about 25% of the time it would work after a dozen retries. other times it would fail completely.

i knew it had something to do with ffmpeg, but upgrading ffmpeg to the latest version made no difference.

painful lessons learned:

  • /live/app/models/media_attachment.rb (line 101) - 'loglevel' => 'fatal' can be changed to 'verbose' to reveal more output from the sidekiq log

  • while it is true you have the latest version of ffmpeg installed, no one promised that mastodon would see it

  • i had a ten-year-old ffmpeg 2.x install lurking in /opt/local that mastodon was using, and completely ignoring my more recent version, all thanks to an errant PATH in my ancient bash_profile 😩

paul,
@paul@oldfriends.live avatar

@arroz

Useless to send to .social admins. They admin and moderate the .social website. (yes, I know, there is some dev crossover, same people, two hats.😉 )

Best, if someone wants to have an issue dealt with to notify devs, is to do so on GitHub where the devs and troubleshooting community are, filed under issues or discussions.

With that said, I am doublechecking my paths, esp old config files. 😂

Thanks for the tag on your post, @vga256

https://github.com/mastodon/mastodon

michael, to random
@michael@thms.uk avatar

Woha, this change to the content retention settings has finally been merged! 🥳

https://github.com/mastodon/mastodon/pull/27733

Thanks @vmstan for your work on this!

rolle, to random
@rolle@mementomori.social avatar

My instance is at version v4.3.0-alpha.3+mementomods-2024-04-06, mastodon.social is at version v4.3.0-nightly.2024-04-18.

"This branch is 124 commits behind"...
Should I keep it... after all, why should I?

(Only MastoAdmin things.)

#MastoAdmin #MementoMoriSocial

GossiTheDog, to random
@GossiTheDog@cyberplace.social avatar

My Mastodon server, cyberplace.social, has received a legal threat in an attempt to have a user's thread deleted. It is styled as a cease and desist.

I have published the email here:
https://github.com/GossiTheDog/Cyberplace/blob/main/LegalThreats/Cease%20and%20Desist%20Order%20-%20Felix%20Juhl

seano, to random
@seano@masto.nyc avatar

Fun fact #2 about masto.nyc hosting:

In the prewar apartment I (and, by association, all the masto.nyc infra) live in there are very few power outlets, so they are all utilized. A few of the outlets are controlled by light switches! Including the outlets for our Internet equipment.

In the early masto.nyc days, guests would mistakenly try to turn off lights but ended up just shutting down the whole site. I 3d printed a little piece so people couldn't flip this switch anymore 💀

paul, to IT
@paul@oldfriends.live avatar

I performed a MTR (Matt's (MY) Traceroute) on my Mastodon S3 storage ( s3.us-east-005.backblazeb2.com) , which is

I used their ingress point for the test. ca005.backblaze.com resolves directly to their ingress points rather than the s3 url.

Any thoughts, on performance, etc. Any comments welcome. I am in a help ticket with one of their Technical Support Engineers.

I'm getting several 443 errors in my Sidekiq jobs, as well as other issues

sb, to mastodon
@sb@fed.sbcloud.cc avatar

Unless you #selfhost your own #mastodon instance, or you're a #mastoAdmin, you have likely taken #fullTextSearch for granted.

Behind search for masto is a resource-hungry #java application called #elasticSearch. Sure you might be able to get Masto to run on a #raspberrypi4, but you won't be running elastic search with it.

Having recently moved my instance to a much more powerful system, I now run ES, and WOW what a difference being able to search post makes! Soooo many times I wanted to reference a toot that had scrolled by but had no way of finding it. Now I do!

Even though I've put some serious resource limits on ES, it's bar none the most resource intensive service running (out of around 25).

Please consider sending a few bucks to the #MastodonAdmin of your instance. Better yet, if you can afford it, sign up for a monthly donation. They need it.

Jeremiah, to privacy
@Jeremiah@alpaca.gold avatar

Why can EU companies not use US-based cloud infrastructure and SaaS products?

Section 702 of the Foreign Intelligence Surveillance Act allows the US government to force US companies to share foreign customer data without a warrant or disclosure.

There is no way to reconcile US govt surveillance with the EU’s Charter of Fundamental Rights on privacy.

More on 702:
https://www.theverge.com/2024/4/20/24135339/fisa-720-reauthorization-senate-lapse-durbin-wyden
https://www.eff.org/pages/upstream-prism

More on EU rulings: https://en.wikipedia.org/wiki/EU%E2%80%93US_Privacy_Shield

Wander, to trustandsafety
@Wander@packmates.org avatar

approaches:
How to deal with accusations and rumors based on outside drama?

This is a tricky one.
Suppose a user is accused of something that is generally frowned upon or worse. However, this has supposedly happened irl, not in your instance, and also the user has a good track record of being active and generating content that doesn't violate the rules at your instance or other instances. There is also no threat to your userbase (as opposed to, for example, in the case of a doxxer).

What is the best approach?:

1- Suspend the user
2- Kick them out (allow them to migrate away, although this just pushes the problem back to the options below)
3- Let them stay and any disagreeing remote instances suspend the user on their side
4- Let them stay and any disagreeing remote instances suspend your whole instance

IMHO, the only sustainable option is option 3, unless it carries a realistic threat to other users. It's not feasible for small instances to be part-time detectives and unless the content posted itself is problematic or there are other reasons, service mentality should prevail if we don't want to push users to super large instances that don't care.

Any concerned remote users or instances are of course free to block on their side.

A whole instance block is not warranted since the instance itself is not spammy or allowing any problematic content, but merely decides not being capable of making a fair assessment of off-platform rumors.

Am I off track?

Jerry, to random
@Jerry@hear-me.social avatar

flipboard.com is a remarkable collection of accounts that are connected to the Fediverse. There are 5,757 accounts from them that are known on this server (we relay with them and get all their content), and these accounts post about every topic imaginable.

You can follow any of their accounts on your favorite Fediverse app! It's great!

I relay with their server so that searches return a richer result set.

Today, I had to create an extra sidekiq process just to handle the surge of posts from them. But, I think it's worth it.

If you have a specific interest, you might find an appropriate account to follow. I'll be looking through their accounts.

Here's a tiny screenshot to show how specialized some accounts are.

thomas, to random
@thomas@metalhead.club avatar

Oh, we've had ~ 12 TB metalhead.club traffic in March. New peak!

#metalheadclub #mastoadmin

CWSmith, to random
@CWSmith@social.mechanizedarmadillo.com avatar

Quick question.

Has anyone connected to @Mastodon been working to switch from Redis since they went closed source? What's going to be involved in moving to a drop in replacement?


OpenTech_AUC, to Amsterdam
@OpenTech_AUC@social.edu.nl avatar

Hello Mastodon! We are a collective of #students and staff from #Amsterdam University College who care about #OpenStandards, #FreeSoftware, and #OpenData.

Every week, we get together to explore, study and organize our way out of the corporate lock-ins that we face at our #university and beyond.

Next to our website, this account will be our main channel to communicate with the world.

We hope that we can connect with others that share our ambitions to push for a free and open digital world.

koen,
@koen@procolix.social avatar

@OpenTech_AUC one of the nice features of running your own instance is that you can increase the number of allowed characters to something a bit more than the default.

That way you can fit those URL's (and your argumentation that has nicely formed sentences instead of abbreviated ones) in your toots and your reactions.

I strongly recommend doing that.

  • All
  • Subscribed
  • Moderated
  • Favorites
  • megavids
  • kavyap
  • DreamBathrooms
  • InstantRegret
  • magazineikmin
  • osvaldo12
  • mdbf
  • Youngstown
  • cisconetworking
  • slotface
  • rosin
  • thenastyranch
  • ngwrru68w68
  • khanakhh
  • JUstTest
  • ethstaker
  • tacticalgear
  • modclub
  • cubers
  • tester
  • everett
  • GTA5RPClips
  • Durango
  • provamag3
  • Leos
  • anitta
  • normalnudes
  • lostlight
  • All magazines