Skip to content
  • Categories
  • Recent
  • Tags
  • Popular
  • World
  • Users
  • Groups
Skins
  • Light
  • Brite
  • Cerulean
  • Cosmo
  • Flatly
  • Journal
  • Litera
  • Lumen
  • Lux
  • Materia
  • Minty
  • Morph
  • Pulse
  • Sandstone
  • Simplex
  • Sketchy
  • Spacelab
  • United
  • Yeti
  • Zephyr
  • Dark
  • Cyborg
  • Darkly
  • Quartz
  • Slate
  • Solar
  • Superhero
  • Vapor

  • Default (No Skin)
  • No Skin
Collapse

NodeBB

  1. Home
  2. General Discussion
  3. So on my ONI instance that I've been use as an alternative fediverse profile for myself for about two years, the full storage used is about 3.4G, but out of that there's 2.5G containing mostly the Delete activities of mastodon.social.

So on my ONI instance that I've been use as an alternative fediverse profile for myself for about two years, the full storage used is about 3.4G, but out of that there's 2.5G containing mostly the Delete activities of mastodon.social.

Scheduled Pinned Locked Moved General Discussion
mastodevfediverseactivitypubdevactivitypub
1 Cross-posts 27 Posts 7 Posters 11 Views
  • Oldest to Newest
  • Newest to Oldest
  • Most Votes
Reply
  • Reply as topic
Log in to reply
This topic has been deleted. Only users with topic management privileges can see it.
  • mariusor@metalhead.clubM This user is from outside of this forum
    mariusor@metalhead.clubM This user is from outside of this forum
    mariusor@metalhead.club
    wrote last edited by
    #1

    So on my ONI instance that I've been use as an alternative fediverse profile for myself for about two years, the full storage used is about 3.4G, but out of that there's 2.5G containing mostly the Delete activities of mastodon.social. Crazy.

    #mastodev #fediverse #ActivityPubDev #ActivityPub

    toddsundsted@epiktistes.comT 1 Reply Last reply
    1
    • ? Guest crossposted this topic to General Discussion
    • mariusor@metalhead.clubM mariusor@metalhead.club

      So on my ONI instance that I've been use as an alternative fediverse profile for myself for about two years, the full storage used is about 3.4G, but out of that there's 2.5G containing mostly the Delete activities of mastodon.social. Crazy.

      #mastodev #fediverse #ActivityPubDev #ActivityPub

      toddsundsted@epiktistes.comT This user is from outside of this forum
      toddsundsted@epiktistes.comT This user is from outside of this forum
      toddsundsted@epiktistes.com
      wrote last edited by
      #2

      @mariusor after reading this, i had to run the same test on my epiktistes/ktistec database. Delete activities are the most numerous activity type, exceeding even Create. 39% of activities are Delete. of those, 45% are from mastodon.social. storage usage is less—those Delete activities are only about 8% of the database size, including indexes.

      mariusor@metalhead.clubM 1 Reply Last reply
      0
      • toddsundsted@epiktistes.comT toddsundsted@epiktistes.com

        @mariusor after reading this, i had to run the same test on my epiktistes/ktistec database. Delete activities are the most numerous activity type, exceeding even Create. 39% of activities are Delete. of those, 45% are from mastodon.social. storage usage is less—those Delete activities are only about 8% of the database size, including indexes.

        mariusor@metalhead.clubM This user is from outside of this forum
        mariusor@metalhead.clubM This user is from outside of this forum
        mariusor@metalhead.club
        wrote last edited by
        #3

        @toddsundsted this scattershot approach to sending Deletes is such a waste of requests and bandwidh on the part of Mastodon.

        There's been so much faff about how verbose ActivityPub is and here we have a trivial optimization to make: just don't send Deletes to the whole network that you have access to. The only interested parties are followers and following collections. Simple!

        #mastoquirks

        nick@hhmx.deN deutrino@mstdn.ioD 2 Replies Last reply
        1
        • mariusor@metalhead.clubM mariusor@metalhead.club

          @toddsundsted this scattershot approach to sending Deletes is such a waste of requests and bandwidh on the part of Mastodon.

          There's been so much faff about how verbose ActivityPub is and here we have a trivial optimization to make: just don't send Deletes to the whole network that you have access to. The only interested parties are followers and following collections. Simple!

          #mastoquirks

          nick@hhmx.deN This user is from outside of this forum
          nick@hhmx.deN This user is from outside of this forum
          nick@hhmx.de
          wrote last edited by
          #4

          @mariusor@metalhead.club

          Unfortunately not so simple.

          Followship may vary over time.
          Accounts could be remotely resolved by fetching conversations.
          Accounts could be remotely resolved manually.
          Accounts could have sent an answer to someone with no relationship.
          And so on.

          @toddsundsted@epiktistes.com

          #Fediverse #ActivityPub #Delete

          toddsundsted@epiktistes.comT mariusor@metalhead.clubM 2 Replies Last reply
          0
          • nick@hhmx.deN nick@hhmx.de

            @mariusor@metalhead.club

            Unfortunately not so simple.

            Followship may vary over time.
            Accounts could be remotely resolved by fetching conversations.
            Accounts could be remotely resolved manually.
            Accounts could have sent an answer to someone with no relationship.
            And so on.

            @toddsundsted@epiktistes.com

            #Fediverse #ActivityPub #Delete

            toddsundsted@epiktistes.comT This user is from outside of this forum
            toddsundsted@epiktistes.comT This user is from outside of this forum
            toddsundsted@epiktistes.com
            wrote last edited by
            #5

            @nick @mariusor ktistec has an ActivityPub "garbage collector". i think i'm just going to tune it to delete older Delete activities. other than archeology, i don't see a ton of value in keeping these around... 🤷

            mariusor@metalhead.clubM 1 Reply Last reply
            0
            • nick@hhmx.deN nick@hhmx.de

              @mariusor@metalhead.club

              Unfortunately not so simple.

              Followship may vary over time.
              Accounts could be remotely resolved by fetching conversations.
              Accounts could be remotely resolved manually.
              Accounts could have sent an answer to someone with no relationship.
              And so on.

              @toddsundsted@epiktistes.com

              #Fediverse #ActivityPub #Delete

              mariusor@metalhead.clubM This user is from outside of this forum
              mariusor@metalhead.clubM This user is from outside of this forum
              mariusor@metalhead.club
              wrote last edited by
              #6

              @nick none of those actions can't be solved by trying to fetch the actor before operating them.

              Also having requests fail should not be an issue for any clients. Receiving 404 and 403 responses is just a normal day on the internet.

              steve@social.technoetic.comS 1 Reply Last reply
              0
              • toddsundsted@epiktistes.comT toddsundsted@epiktistes.com

                @nick @mariusor ktistec has an ActivityPub "garbage collector". i think i'm just going to tune it to delete older Delete activities. other than archeology, i don't see a ton of value in keeping these around... 🤷

                mariusor@metalhead.clubM This user is from outside of this forum
                mariusor@metalhead.clubM This user is from outside of this forum
                mariusor@metalhead.club
                wrote last edited by
                #7

                @toddsundsted I'm considering adding extra logic to the Delete workflow so if the Deleted object does not exist locally (previously fetched or created) the Delete doesn't get processed or persisted...

                @nick

                1 Reply Last reply
                0
                • mariusor@metalhead.clubM mariusor@metalhead.club

                  @toddsundsted this scattershot approach to sending Deletes is such a waste of requests and bandwidh on the part of Mastodon.

                  There's been so much faff about how verbose ActivityPub is and here we have a trivial optimization to make: just don't send Deletes to the whole network that you have access to. The only interested parties are followers and following collections. Simple!

                  #mastoquirks

                  deutrino@mstdn.ioD This user is from outside of this forum
                  deutrino@mstdn.ioD This user is from outside of this forum
                  deutrino@mstdn.io
                  wrote last edited by
                  #8

                  @mariusor @toddsundsted Mastodon hasn't ever been known for being efficient or elegant, and after being here for a decade I don't see that ever changing

                  mariusor@metalhead.clubM 1 Reply Last reply
                  0
                  • deutrino@mstdn.ioD deutrino@mstdn.io

                    @mariusor @toddsundsted Mastodon hasn't ever been known for being efficient or elegant, and after being here for a decade I don't see that ever changing

                    mariusor@metalhead.clubM This user is from outside of this forum
                    mariusor@metalhead.clubM This user is from outside of this forum
                    mariusor@metalhead.club
                    wrote last edited by
                    #9

                    @deutrino I know, I know, I'm the biggest complainer about Mastodon's lack of effort when it comes to better support the specification. I suspect everyone in their team has me on mute already. 😄

                    @toddsundsted

                    1 Reply Last reply
                    0
                    • mariusor@metalhead.clubM mariusor@metalhead.club

                      @nick none of those actions can't be solved by trying to fetch the actor before operating them.

                      Also having requests fail should not be an issue for any clients. Receiving 404 and 403 responses is just a normal day on the internet.

                      steve@social.technoetic.comS This user is from outside of this forum
                      steve@social.technoetic.comS This user is from outside of this forum
                      steve@social.technoetic.com
                      wrote last edited by
                      #10

                      @mariusor @nick How does fetching the actor's current state help with knowing the original distribution of an activity? What happens when a popular follower boosts a post to thousands of their followers and to many servers (and their followers re-boost to yet more servers, etc.). There are also relays that may have forwarded an activity to many servers and is now no longer active to forward Deletes. If your trivial solution is effective, I'd like to see a full description/analysis in an FEP.

                      mariusor@metalhead.clubM 1 Reply Last reply
                      0
                      • steve@social.technoetic.comS steve@social.technoetic.com

                        @mariusor @nick How does fetching the actor's current state help with knowing the original distribution of an activity? What happens when a popular follower boosts a post to thousands of their followers and to many servers (and their followers re-boost to yet more servers, etc.). There are also relays that may have forwarded an activity to many servers and is now no longer active to forward Deletes. If your trivial solution is effective, I'd like to see a full description/analysis in an FEP.

                        mariusor@metalhead.clubM This user is from outside of this forum
                        mariusor@metalhead.clubM This user is from outside of this forum
                        mariusor@metalhead.club
                        wrote last edited by
                        #11

                        @steve I'm not sure I understand your question. What does a popular account posting have to do with the dissemination list of Delete activities? Are you suggesting that the problem is that they're sending to deleted actors? Perfect, now the server knows to stop sending there.

                        @nick

                        steve@social.technoetic.comS 1 Reply Last reply
                        0
                        • mariusor@metalhead.clubM mariusor@metalhead.club

                          @steve I'm not sure I understand your question. What does a popular account posting have to do with the dissemination list of Delete activities? Are you suggesting that the problem is that they're sending to deleted actors? Perfect, now the server knows to stop sending there.

                          @nick

                          steve@social.technoetic.comS This user is from outside of this forum
                          steve@social.technoetic.comS This user is from outside of this forum
                          steve@social.technoetic.com
                          wrote last edited by
                          #12

                          @mariusor @nick No, not suggesting that at all. *Boosts* will cause an activity to be be fetched by servers that aren't directly known to the original actor. (I have no idea what you're talking about wrt deleted actors.)

                          steve@social.technoetic.comS mariusor@metalhead.clubM 2 Replies Last reply
                          0
                          • steve@social.technoetic.comS steve@social.technoetic.com

                            @mariusor @nick No, not suggesting that at all. *Boosts* will cause an activity to be be fetched by servers that aren't directly known to the original actor. (I have no idea what you're talking about wrt deleted actors.)

                            steve@social.technoetic.comS This user is from outside of this forum
                            steve@social.technoetic.comS This user is from outside of this forum
                            steve@social.technoetic.com
                            wrote last edited by
                            #13

                            @mariusor @nick AFAICT, Fediverse Delete is practically going to be "best effort" and not complete. The question is how much effort is an implementation willing to make. Only sending a Delete to followers and following collections seems to me to be minimal effort with minimal effectiveness. On the high effort side, an implementation could track every domain that's ever fetched a specific document and send Deletes to those. However, that's doesn't seem practical.

                            mariusor@metalhead.clubM 1 Reply Last reply
                            0
                            • steve@social.technoetic.comS steve@social.technoetic.com

                              @mariusor @nick No, not suggesting that at all. *Boosts* will cause an activity to be be fetched by servers that aren't directly known to the original actor. (I have no idea what you're talking about wrt deleted actors.)

                              mariusor@metalhead.clubM This user is from outside of this forum
                              mariusor@metalhead.clubM This user is from outside of this forum
                              mariusor@metalhead.club
                              wrote last edited by
                              #14

                              @steve then you should read my top post, to which
                              @nick was replying.

                              steve@social.technoetic.comS 1 Reply Last reply
                              0
                              • steve@social.technoetic.comS steve@social.technoetic.com

                                @mariusor @nick AFAICT, Fediverse Delete is practically going to be "best effort" and not complete. The question is how much effort is an implementation willing to make. Only sending a Delete to followers and following collections seems to me to be minimal effort with minimal effectiveness. On the high effort side, an implementation could track every domain that's ever fetched a specific document and send Deletes to those. However, that's doesn't seem practical.

                                mariusor@metalhead.clubM This user is from outside of this forum
                                mariusor@metalhead.clubM This user is from outside of this forum
                                mariusor@metalhead.club
                                wrote last edited by
                                #15

                                @steve so between impractical, minimal effectiveness, and scattershot wasting everyone's resources and time, which one is better in your opinion?

                                @nick

                                steve@social.technoetic.comS 1 Reply Last reply
                                0
                                • mariusor@metalhead.clubM mariusor@metalhead.club

                                  @steve then you should read my top post, to which
                                  @nick was replying.

                                  steve@social.technoetic.comS This user is from outside of this forum
                                  steve@social.technoetic.comS This user is from outside of this forum
                                  steve@social.technoetic.com
                                  wrote last edited by
                                  #16

                                  @mariusor @nick I had already read that. You discuss "Deletes", in general, but maybe you only meant "actor deletes"?

                                  mariusor@metalhead.clubM 1 Reply Last reply
                                  0
                                  • mariusor@metalhead.clubM mariusor@metalhead.club

                                    @steve so between impractical, minimal effectiveness, and scattershot wasting everyone's resources and time, which one is better in your opinion?

                                    @nick

                                    steve@social.technoetic.comS This user is from outside of this forum
                                    steve@social.technoetic.comS This user is from outside of this forum
                                    steve@social.technoetic.com
                                    wrote last edited by
                                    #17

                                    @mariusor @nick "Scattershot and wasting everyone's time" is an opinion. Some people may appreciate the extra effort to make Deletes more effective. I'd hope an implementation that makes minimal effort to make Delete effective would clearly disclose that to admins and users.

                                    mariusor@metalhead.clubM 1 Reply Last reply
                                    0
                                    • steve@social.technoetic.comS steve@social.technoetic.com

                                      @mariusor @nick "Scattershot and wasting everyone's time" is an opinion. Some people may appreciate the extra effort to make Deletes more effective. I'd hope an implementation that makes minimal effort to make Delete effective would clearly disclose that to admins and users.

                                      mariusor@metalhead.clubM This user is from outside of this forum
                                      mariusor@metalhead.clubM This user is from outside of this forum
                                      mariusor@metalhead.club
                                      wrote last edited by
                                      #18

                                      @steve 2.5GB of activitypub data out of a total of 3.4GB is not waste when I haven't seen the Creates that generated those Deleted entities?

                                      @nick

                                      steve@social.technoetic.comS 1 Reply Last reply
                                      0
                                      • steve@social.technoetic.comS steve@social.technoetic.com

                                        @mariusor @nick I had already read that. You discuss "Deletes", in general, but maybe you only meant "actor deletes"?

                                        mariusor@metalhead.clubM This user is from outside of this forum
                                        mariusor@metalhead.clubM This user is from outside of this forum
                                        mariusor@metalhead.club
                                        wrote last edited by
                                        #19

                                        @steve yes, the actor ones are the most egregious, because additionally to just wasting disk space, they waste resources when the instances try to dereference the actors that own the HTTP-Signature that signed the Delete and find them gone.

                                        @nick

                                        1 Reply Last reply
                                        0
                                        • mariusor@metalhead.clubM mariusor@metalhead.club

                                          @steve 2.5GB of activitypub data out of a total of 3.4GB is not waste when I haven't seen the Creates that generated those Deleted entities?

                                          @nick

                                          steve@social.technoetic.comS This user is from outside of this forum
                                          steve@social.technoetic.comS This user is from outside of this forum
                                          steve@social.technoetic.com
                                          wrote last edited by
                                          #20

                                          @mariusor @nick I don't know. See my comments about *boosting* where you may have cached entity without receiving a Create. Nick mentioned other examples (manual activity fetch, etc.).

                                          hongminhee@hollo.socialH mariusor@metalhead.clubM 2 Replies Last reply
                                          0

                                          Hello! It looks like you're interested in this conversation, but you don't have an account yet.

                                          Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.

                                          With your input, this post could be even better 💗

                                          Register Login
                                          Reply
                                          • Reply as topic
                                          Log in to reply
                                          • Oldest to Newest
                                          • Newest to Oldest
                                          • Most Votes


                                          • Login

                                          • Don't have an account? Register

                                          • Login or register to search.
                                          Powered by NodeBB Contributors
                                          • First post
                                            Last post
                                          0
                                          • Categories
                                          • Recent
                                          • Tags
                                          • Popular
                                          • World
                                          • Users
                                          • Groups