• 0 Posts
  • 16 Comments
Joined 1 year ago
cake
Cake day: June 16th, 2023

help-circle



  • This is the correct answer, IMO.

    I loved using XMPP back in the day, but I struggled talking with people who weren’t on the same server as me because of spec and client variations.

    While Synapse is a resource hog, it (and Element) - to a certain degree - does the job. Can’t wait until sync v3 lands in the main server.

    The only issue I have is with one friend who insists on deploying his own version of Synapse, but can’t figure out coturn and - as a result - we can’t voice chat properly.

    Goddammit. Two steps forward, one step backward. 😅






  • Sorry man, I don’t know what to tell you. I’ve got a pretty medium end VPS on which I host my Matrix instance - only had to add an extension for storage after the first few years when the DB got too big. Things were never as bad as you said early on, and as time passed I absolutely got to the point where it would take 10-20 seconds to sync - but this was after 2 years or so of constant use.

    The reason why it takes long is because of the size of the sync payload - logically, for a new server/user, this really shouldn’t be that big (unless you’re in rooms like Matrix HQ). So, genuinely, look into optimization: postgres, your web server (nginx, apache, caddy), and limiting your users from accessing “problematic” rooms.

    Barring that just deploy the sliding sync proxy and be done with it. It’s not really a problem that requires you to attempt it a thousand times.

    So either you put some fancy wizardry into your system or you’re just in denial.

    It’s called pure Debian, baby. Also, you’ll need a decent chunk of RAM if you don’t have that yet. Avoid a pagefile if you can.






  • As a person who oversaw the implementation of GDPR in a large software house (which wasn’t EU specific, but had to in order to operate legally in the EU), the requirements were:

    1. Allow users to request data deletion or a copy of their data.
    2. If the former, delete all data of their data on the server, send it to them, and then (this was the important part) forward the data deletion request to every single partner we were working with.

    For us, this was multiple ad companies. We had to e-mail each one, ask them about their GDPR implementation (most of them were somewhere between “we’re thinking about it” and “we have an e-mail address you can send something automated to and we’ll get to it sometime within the next month”), and then build an automated back-end system to either query their APIs for automated deletion, or craft/send e-mails for the more primitive companies.

    As far as the data being deleted, it was anonymized IDs that were tied to their advertising IDs from their mobile phones. I used to try and argue that “no, it’s anonymous” - but we also had some player data (these were games) associated with that, so we ended up just clearing house and deleting everything on request.

    So, legally, this means every instance - in order to be GDPR compliant - would have to inform every instance it federates with that a user wants their data deleted. If you’re not doing that, you’re not fully compliant.

    Kind of shitty, but that’s how it went for me. (this was back when GDPR was first being released)

    Edit: Also, the one month thing was relevant: you have 30 days to delete GDPR stuff after receiving a data clear request. I don’t recall what the time was for a “see my data” request. Presumably, though, on Lemmy the latter is superfluous as all your data is already present on your profile page. An account export option would be enough to satisfy that.