Conversation

Notices

1. FWC (fuckingwhocaresdude@5dollah.click)'s status on Sunday, 19-Nov-2023 01:53:46 JST FWC
   Still a minor delay in outgoing federation. Anyone know anything about optimizing the beam for Pleroma?
   •  (mint@ryona.agency)'s status on Sunday, 19-Nov-2023 01:53:44 JST 

     in reply to

     • FrailLeaf
     @FrailLeaf @FUCKINGWHOCARESDUDE No idea. From my experience, delays are mostly caused by large federation queue, one can either wait it out or drop the queue in entirety, but that might result in some posts, remote or local, never federating until they get interacted with.
   • FrailLeaf (frailleaf@ryona.agency)'s status on Sunday, 19-Nov-2023 01:53:45 JST FrailLeaf

     in reply to

     • 
     @FUCKINGWHOCARESDUDE cc @mint
   •  (mint@ryona.agency)'s status on Sunday, 19-Nov-2023 03:31:38 JST 

     in reply to

     • FrailLeaf
     @FUCKINGWHOCARESDUDE @FrailLeaf OTP releases bundle Erlang/Elixir with them, so the instruction is definely about the source releases. Looks like vm.args exists ( ./releases/2.6.50-0-geef9a89c4-ryona-dev/vm.args, change the version to your own) and is read by starting script (./bin/pleroma) so it might be as easy as adding your settings there, but I haven't done it as I had no need to.
   • FWC (fuckingwhocaresdude@5dollah.click)'s status on Sunday, 19-Nov-2023 03:31:39 JST FWC

     in reply to

     • 
     • FrailLeaf
     @mint @FrailLeaf Do you know about optimizing the beam as below in the docs? I have no clue if this applies to the OTP installation or what. It also doesn't say what needs to be a systemd service or if the instance needs to be stopped here. I understand if you don't know. I certainly don't.
     
     https://docs.pleroma.social/backend/configuration/optimizing_beam/
   • FWC (fuckingwhocaresdude@5dollah.click)'s status on Sunday, 19-Nov-2023 04:22:37 JST FWC

     in reply to

     • 
     • FrailLeaf
     @mint @FrailLeaf Thanks, will try that! Anything else you know about that could cause federation to be slow?
      likes this.
   •  (mint@ryona.agency)'s status on Sunday, 19-Nov-2023 04:23:19 JST 

     in reply to

     • FrailLeaf
     @FUCKINGWHOCARESDUDE @FrailLeaf Other than possible network issues, no idea. What are your servers specs, by the way?
   • FWC (fuckingwhocaresdude@5dollah.click)'s status on Sunday, 19-Nov-2023 04:29:53 JST FWC

     in reply to

     • 
     • FrailLeaf
     @mint @FrailLeaf it's running on an HP HP Compaq 6200 Pro Desktop with an i5-2300 and 14 Gb of RAM
      likes this.
   •  (mint@ryona.agency)'s status on Sunday, 19-Nov-2023 04:31:41 JST 

     in reply to

     • FrailLeaf
     @FUCKINGWHOCARESDUDE @FrailLeaf So a dedicated hardware, like mine (A8-6500B, 24 gigs). Beam shouldn't be an issue, since its troubleshooting mostly applies to instances living on overprovisioned VPS hosts.
   • FWC (fuckingwhocaresdude@5dollah.click)'s status on Sunday, 19-Nov-2023 04:38:33 JST FWC

     in reply to

     • 
     • FrailLeaf
     @mint @FrailLeaf It seems like there's a way to optimize it for dedicated hardware, though. The flip side is that I honestly don't know what the issue is, so that makes this a total shot in the dark.
      likes this.
   •  (mint@ryona.agency)'s status on Sunday, 19-Nov-2023 04:47:17 JST 

     in reply to

     • FrailLeaf
     @FUCKINGWHOCARESDUDE @FrailLeaf Try playing around with federation pool sizes.
     https://git.pleroma.social/pleroma/pleroma/-/blob/develop/config/config.exs#L803