Conversation

Notices

1. Embed this notice
   nyanide :nyancat_rainbow::nyancat_body::nyancat_face: (nyanide@lab.nyanide.com)'s status on Sunday, 24-Aug-2025 12:45:33 JST nyanide :nyancat_rainbow::nyancat_body::nyancat_face:
   Dual Pleroma is hard, triple Pleroma even harder. I dunno what to do about the Tor instance, I am unsure if I want to keep it running in my bedroom indefinitely.
   • Embed this notice
     prettygood (prettygood@socially.drinkingatmy.computer)'s status on Sunday, 24-Aug-2025 12:49:21 JST prettygood

     in reply to
     @nyanide I thought about trying to run two of them behind an nginx proxy. I think I could just set the web.endpoint to another port number on the second instance, yeah?
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: likes this.
   • Embed this notice
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: (nyanide@lab.nyanide.com)'s status on Sunday, 24-Aug-2025 12:52:34 JST nyanide :nyancat_rainbow::nyancat_body::nyancat_face:

     in reply to

     • prettygood
     @prettygood Yes you can do that but Elixir will give in after a while if you decide to run too many Pleroma servers at once. Why? Who the fuck knows.
     prettygood likes this.
   • Embed this notice
     prettygood (prettygood@socially.drinkingatmy.computer)'s status on Sunday, 24-Aug-2025 12:53:15 JST prettygood

     in reply to
     @nyanide fair enough I also give in when too much is expected from me so that's very relatable behavior.
   • Embed this notice
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: (nyanide@lab.nyanide.com)'s status on Sunday, 24-Aug-2025 12:56:56 JST nyanide :nyancat_rainbow::nyancat_body::nyancat_face:

     in reply to

     • prettygood
     @prettygood There's also Postgres taking up too much storage space if you start to run more than 1 Pleroma instance at once. FRE literally sat doing basically nothing for months and DB dump was 7 gigs, Pleroma is a hungry girlie.
     prettygood likes this.
   • Embed this notice
     prettygood (prettygood@socially.drinkingatmy.computer)'s status on Sunday, 24-Aug-2025 13:03:41 JST prettygood

     in reply to
     @nyanide hmm, yeah, I hadn't considered that. Federating be like that.
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: likes this.
   • Embed this notice
     tsoifan1997 (sysrq@lab.nyanide.com)'s status on Sunday, 24-Aug-2025 13:06:28 JST tsoifan1997

     in reply to
     @nyanide But AU Falcon FUNNY
   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Sunday, 24-Aug-2025 15:52:19 JST Phantasm

     in reply to

     • prettygood
     @nyanide @prettygood I've seen beam get hungry for memory for no reason, oom and be fine for weeks again multiple times. It went from maybe a gigabyte of allocated memory to 4x of that in a minute.
     
     I don't think it likes running with other applications on the same host like with the process spin trick it uses to convince the kernel scheduler to not pause the process and give the resources to something else. There might be some obscure argument for it, but who the fuck knows those.
     prettygood and nyanide :nyancat_rainbow::nyancat_body::nyancat_face: like this.
   • Embed this notice
     prettygood (prettygood@socially.drinkingatmy.computer)'s status on Sunday, 24-Aug-2025 15:53:18 JST prettygood

     in reply to

     • Phantasm
     @phnt @nyanide in my experience the only problem I've had has been epmd not exiting gracefully and needing to be killed manually. I suspect this is because both pleroma and ejabberd are using it so when one service stops it doesn't accept the sighup or whatever because the other thing is using it. So I just added "pkill epmd" to my scripts :-p
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: likes this.
   • Embed this notice
     Phantasm (phnt@fluffytail.org)'s status on Sunday, 24-Aug-2025 16:11:45 JST Phantasm

     in reply to

     • prettygood
     @prettygood @nyanide epmd is supposed to be started as a separate service since it is supposed to be a nameserver for other Erlang nodes on the network, but very few people do that. I think only Debian has a service file for it.
     prettygood and nyanide :nyancat_rainbow::nyancat_body::nyancat_face: like this.
   • Embed this notice
      (mint@ryona.agency)'s status on Sunday, 24-Aug-2025 16:22:25 JST 

     in reply to
     @nyanide Just put it in lxc container or something.
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: likes this.
   • Embed this notice
     :blank: (i@declin.eu)'s status on Monday, 25-Aug-2025 00:13:13 JST :blank:

     in reply to

     • prettygood
     • Phantasm
     @phnt @nyanide @prettygood it's in the docs https://docs.pleroma.social/backend/configuration/optimizing_beam/#virtual-machine-andor-few-cpu-cores
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: likes this.
   • Embed this notice
     :blank: (i@declin.eu)'s status on Monday, 25-Aug-2025 00:16:24 JST :blank:

     in reply to
     @nyanide this is why my vaporware runs an infinite amount of domains at the moment, wonder how much hassle a multi-tenant pleroma would be, a lot of it sucks because you'd still be duplicating most of the data among prefixes
     nyanide :nyancat_rainbow::nyancat_body::nyancat_face: likes this.