Posts made by doodlemania2

@scooke yep absolutely!

@alwynispat A lot of the fediverse apps are being tried by different appdev folks (myself included) - there are some challenges getting some (like lemmy et al) into a single container, but progress is happening! I helped with Pixelfed and Mastodon a bit. Now working on writefreely, lemmy, and bookwyrm as time permits

@jdaviescoates filesystem to local did it!

@jdaviescoates I'm stuck on the URL still being munged with half local and half s3 with the profile pic...

@jdaviescoates yep - it's a bit of an odd situation. and blocking me from going live

@jdaviescoates similar impact on the latest build since we switched to dev branch! cc: @girish FYI only. Hope they get that bug fixed

@robi slow and steady!

@jdaviescoates tried again on another box, same story - it's early as @girish said but just poking on things!

@jdaviescoates Yeah, not sure how to move it

@girish That's amazing! Sorry I wasn't able to get to this, been focused on the other stuffs of late -

Quick question - on initial install, I'm stuck at /installer/index:

Clicking next just returns to that page.
Seems like when I hit Next, it bounces the app:

@alwynispat you've got DB_POOL specified twice
I'm doing:

@nichu42 I didn't do the cache bit which can't easily be done in cloudron, media loads directly from "s3" (mine is also in idrive). It's definitely slower and you SHOULD use a cache (in or very near to your cloudron server) if you have more than a few dozen users.

I was thinking about packaging up a VERY simple custom cloudron app that did exactly that but figured my scenario was too esoteric.

When you put the details into your .env file, then restart and all new media will go to "s3" but you have to repopulate your media for historical stuff in your bucket with tootctl - but you'll be able to rm -rf the cache folder on your vps!

@jdaviescoates I moved mine based on this article: https://stanislas.blog/2018/05/moving-mastodon-media-files-to-wasabi-object-storage/#setting-up-a-nginx-reverse-proxy-with-cache-for-the-bucket

@nichu42 best practice would be to offload media entirely to cheaper blob endpoint, as opposed to expensive block storage in your VM

@archos I think the automated system only picks up when there's a 'release' that dansup makes.

@d19dotca I also noticed an increase in performance. One thing I did notice though - my public buckets were part of the beta and included anonymous root "viewing" of sorts - it generated an XML file of the contents of each public bucket. A quick email to support and they had me turn the bucket to private, then back to public to correct. Something to check if you or someone you love may be an early public bucket adopter there!

@d19dotca This is cool! How do I "implement" it?

that did the trick nicely!

@jdaviescoates ohhhh! doinks - thanks sir

I think I may have gotten to a threshold of sorts. The sidekiq running in my Mastodon instance isn't able to keep up. Top shows it running in the container no problem, but apparently I need more than one instance to get some parallelism rocking. Top also showed the container wasn't pegged at 100% CPU, so it had room to grow. I also bumped up the Resources from default 50% to 75%.

So, I ended up with about a 6 hr backlog (around 40,000 messages) in my queues. In order to get them to clear, I opened up two terminal instances and manually ran bundle exec sidekiq in each and that chewed through them pretty quickly. Once the queues were drained, I closed them out but it started building back up again - the default sidekiq is definitely running, just a lot slower than needed.

So, wondering if we could have a config option to ramp additional sidekiqs up?

@humptydumpty same as you - noco felt too "fluffy"

@timconsidine went with 6 and it fired right up - may try 4 to start

well I feel dumb lol - sorry about that

let me try that - did not think it was memory!

@privsec nextcloud talk HP backend is $4/domain/mo

@jdaviescoates oh yes, you can see all my goodies at https://www.thedoodleproject.com

I should add - my existing baserow (up to date) runs fine. Maybe something went wonky in the new install routine?

Can anyone try to spin up a Baserow and see if it successfully starts? Have tried on two different Cloudrons and am getting similar:

@marcusquinn awe thanks !!!

Hey all, I just obtained my LLC status for The Doodle Project! If you're looking for assistance in Cloudron help or are interested in some of my services I offer, <insert shameless plug here>. Delighted to be working with several of you on hosting Nextcloud Talk High Performance Backend, Big Blue Button (both of those offered in both US and EU), as well as a bunch of other shiny fun things. Holler if you need anything!

Of course, if you have questions about the LLC'ness, happy to answer any time. Especially important if you're considering setting up shop in the US with a Mastodon (or similar) instance to provide coverage for the ugliness out there. Have also been thinking about offering a service to help customers host their own instances of that underneath my 'umbrella' - let me know if interested.

See Twitter today, apparently Substack is (at least in part), now running on Ghost!

@shanelord01 I don't think any progress was made. Still an open conversation.

I've been running Lokinet for some period of time. Love it for all the reasons mentioned by others. I think it might be as simple as binding CR to the network device created by Lokinet. I've always wanted to try it, but been too scared

@LoudLemur Same could be asked for something like Lokinet. Overlay network support in general would be cool

@AyoubHabra Apologies for the delay - somehow I missed the alert on this. I've not changed anything to permit/reject traffic - not sure why it works for me but not you That said, I haven't tried to formally package it for the ecosystem due to lack of interest, but will take a peak and see if I can figure out why it's not working for you.

@girish you still have / making use of your Azure credits from a few years back?

Yeah, the same doc shows how to set up an Ngnix cache - I'm gonna try to hack something together to see if I can front that in CR somehow. Could serve some other generic purposes too for other systems that could benefit from an S3 cache.

@jdaviescoates I can +1 the need for a cache - moving to S3 works great but it does slow things way down. I am trying to find a way to do a local nginx cache with CR but coming up short so far.

@robi yeah, I tried to cat / tail it, was unicode garbage. me thinks corrupted

@girish well, the thing that kept getting overwhelmed was netns ip6 and i also noticed that my tailscale agent was running at like 4.6GB in memory, so thought, okay, let me ping something. it took a solid 40 seconds to even return a ping but it was correct. so that's when i looked at hosts.

did a cat to /dev/null on it and boom, all back to good. i am thinking that some dns thing went haywire - but I don't think it's CR's fault cause I've never seen CR try to write to hosts. Will dig a bit more but I also have an LTS upgrade due, so perhaps there's a bug someplace in Ubuntu with bind.

@girish okay - I think I fixed it - the /etc/hosts file was almost 40GB in size of garbage (shrug)
I emptied it out and replaced it with just default localhost items and restarted everything. MUCH happier it is.
Thank you so much for looking at it with me - it was in a very bad way!

Now to see if I can figure out why hosts file got gummed up

@girish sorry for delay, got sidetracked on another thing - docker disabled and restarted!

@RazielKanos Although having trouble at this very moment, I don't think it's there fault, I REALLY like ssdnodes.com. They're always running good sales. Here's my referral link if you're interested

https://www.ssdnodes.com/manage/aff.php?aff=1042

@doodlemania2 @doodlenode:~$ sudo cloudron-support --enable-ssh
[sudo] password for derek:
Enabling ssh access for the Cloudron support team...Done

@girish sure thing - will enable from terminal!

@girish haha, nah, ssd'nodes that @robi introduced me to - they're just darn close to amazing!

They are still looking but I am starting to wonder if docker is just corrupted somehow.

@robi enabled and restarted - still remain stuck - still digging, wondering if docker somehow went wonky. VPS team also looking.

@girish yeah, it's just running really hot this type of command over and over:
netns-create/var/run/docker/netns/f4e6ea807391

htop gives a slightly better visual:

Something about set-ipv6 maybe? (Not using IPv6 on this CR I don't think)

This morning my CR ground to a halt. I thought it was getting VPS bound since I've been "popular" the past few days, but they said all good so I kept digging. In the VM, there's a bin running REALLY hot. In top it shows up as "exe" which made me worry a bit:

1563 root 20 0 4926308 2.9g 28356 R 102.3 6.2 1:07.44 exe

So I ps-ax'd the PID and it came back as
1563 ? Rl 0:26 set-ipv6 /var/run/docker/netns/3939c4953582 all false

After a bit, that PID goes away and is replaced by another PID just like it. VM is up but CR is down.

Did a reboot, still persists. Any ideas?

Yeah, at the moment it's a script that calls the mastodon api to add a git-ops derived block list. Hachyderm has one too. It's a neat idea (philosophical questions not withstanding).

I think the easiest approach would be to download the code locally and have it run against your Mastodon instance and that would allow folks to choose to use it if they want or not, but not bake it into the default app.

But I may have read the instructions wrong!

@RazielKanos I can take another look - hopefully it has progressed some!

I ran some traps on this thread posted by a law group that had set up their own instance. Also @girish echo'ed the DCMA business in another thread. Thought I'd share:

"Do Future You a huge favor, mitigate your potential liability, and register with the copyright office and designate an agent to receive DMCA reports right now. https://copyright.gov/dmca-directory/

It costs $6 a year. Use a forwarding phone # and a PO Box or other address you check frequently, not your home address: it will be public. Without this registration, you WILL be held liable for any copyright violation on your instance.

I also strongly recommend that if you run an instance, incorporate as a LLC and -- this is the critical part -- take out an umbrella liability insurance policy with coverage of at least $2m per incident and covers attorney costs and fees.
(If you have homeowners or renters' insurance, and you should, this should be a very cheap rider on the policy. 99% of the time it will be unnecessary and the other 1% it will save your ass completely.)

I am somehow not surprised that only two Mastodon instances have registered designated agents and one of them is the Lawprofs instance.

<tweak by Derek>Also: "I found someone on my instance posting CSAM" is not the time to have to learn the reporting process for that vileness. Learn and report ASAP via NCMEC in case of CSAM posted to your service. Legally mandated.</tweak by Derek>

But yeah, for copyright, feel free to steal our DMCA policy, it's CC-BY-SA. (Note that we willingly accept a level of potential risk around repeat offender account termination that may be outside your risk tolerance.) https://dreamwidth.org/legal/dmca

Our ToS and privacy policy are also CC, and you're welcome to use them, but they should be more customized for your individual situation than just "take this and change it to your contact info".

The relevant section of copyright law is 17 USC §512 and you can read it here: https://law.cornell.edu/uscode/text/17/512 You are protected from liability for your users' copyright violations, but ONLY if you follow this process, and the process includes registering a designated agent.

I can't give you legal advice and none of this is intended as such, but if you're absolutely at sea and have never heard of any of this before, I can try to answer general questions.

I have posted the "guide to familiarize you with various legal obligations involved in hosting a server that accepts user-generated content such as a Mastodon instance"! It's here: https://denise.dreamwidth.org/91757.html

Original thread: https://twitter.com/rahaeli/status/1593819064161665024 Tweaked by me for clarity.

Net - if you're running an instance that's open - be aware. If you're running a closed or single user instance, still not a terrible idea. I'm going to be doing it myself for my single user instance because I also run an open relay to help folks out.

@shanelord01 (waves) it auto-approves after a few minutes assuming your side is set up cleanly - I see you and welcome!

Many people have tried to relay over the past few weeks. Some don't have SSL (hard pass), some are using fake FQDN (hard pass), and then of course, we had a few bad actors come along and I had to boot em.

Be mindful of your blocks and mutes - whole instance blocking is a big sledgehammer unless it's just a whole instance of bad. Also, if you see someone on my relay and you're like, ewe, Derek should know about that, please ping me (at)derek(at)toot.thedoodleproject.net - I am watching but it's a LOT to watch and any visibility on bad instances is appreciated - let me know and I'll block em at the relay.

Just an FYI - I moved cache off local to an S3 endpoint. It was cake. It's a bit 'slower' now but you can also add a local'ish cache to speed things up a bit.

@scooke So I heard back from email based support - they turned public buckets on for my account!

went with their main tech support email

@scooke anyone know a contact? Would like to stay with them but no public buckets...lol - maybe they can bit flip it for me?

@scooke funny story - i just migrated all my stuff to idrive for my mastodon instance only to discover - twoops - they don't appear to allow public buckets!?

@redegelde may be a few steps between there and 4.0, but yes, just keep upgrading (don't forget to backup!)

@d19dotca I keep thinking that between these newcomers with their free egress and cloudflare shaming them, that Azure and AWS et al will drop the egress fees (or at least lower them)!