Conversation

Notices

1. Embed this notice
   Puniko ? (puniko@mk.absturztau.be)'s status on Wednesday, 19-Oct-2022 18:23:55 JST Puniko ?

   https://polaroid.absturztau.be is now running on latest pixelfed

   • Embed this notice
     Puniko ? (puniko@mk.absturztau.be)'s status on Wednesday, 19-Oct-2022 18:29:09 JST Puniko ?

     in reply to

     • Vale@kujike.nai

     @valerauko@pawoo.net it mostly depends on how many users, what software specifically and stuff. most microblogging instances, that are small in user size, can be run on a fairly cheap vps afaik

   • Embed this notice
     Vale@kujike.nai (valerauko@pawoo.net)'s status on Wednesday, 19-Oct-2022 18:29:10 JST Vale@kujike.nai

     in reply to

     @puniko just for reference, what are the resource (memory, cpu, space) requirements of the various fedi stuff hosted?

   • Embed this notice
     Vale@kujike.nai (valerauko@pawoo.net)'s status on Wednesday, 19-Oct-2022 18:46:50 JST Vale@kujike.nai

     in reply to

     @puniko what's your experience with the stuff on absturztau.be?

   • Embed this notice
     Puniko ? (puniko@mk.absturztau.be)'s status on Wednesday, 19-Oct-2022 18:46:50 JST Puniko ?

     in reply to

     • Vale@kujike.nai

     @valerauko@pawoo.net well. i started out with fedi.absturztau.be on a proxmox. did run ok, only issue was, that the proxmox cluster had a broken cephfs config which meant that disk IO was really really poor. if it weren't for that, it would have run fine. the container i used used 2 cores and had like 4 or 8G or RAM initally (i don't remember) which was plenty for starting out.
     
     my instance got bigger tho and the disk io issue got worse and worse. so i bugged my boss about when dedicated hardware is gonna be phased out of the company so i could reuse that hardware. after like 3 months or so, i got myself a dedicated server, quite beefy. since then, i got plenty of cpu, ram and stuff left for running a bunch of services (it only coughs up on itself sometimes when a lot is going on, which usually just means that the response time is higher than usual).
     
     so to summarize: enough cores for each service, enough ram to handle the service + database and quick disks are kinda mandatory for running mutliple services at once