As the question states. I have the skiils and know how to setup a new instsnce and from looking through the documentation its relatively straight forward. Im just wondering would it actually help the fediverse and lemmy as a whole? There seems to already be plenty of instances to choose from. What are your guys thoughts?
@bionicspud, I’ve been thinking of doing the same thing for the same reasons. How much storage does your instance use? I haven’t looked in to how asset caching works yet. It seems like it could easily get out of hand.
Running the Instance for 3 Days now. Added multiple new Subs and cleared pictrs (but not Postgres) folder yesterday. My instance is mostly used by me.
87M ./pictrs
335M ./postgres
Edit: Small update. Original Post was posted ~9AM. Now its ~11AM
112M ./pictrs
351M ./postgres
As far as i know, images shouldn’t be stored locally if not posted from the local instance. This looks to be a proxing problem as far as i understood.
what happens when you clear pictrs? you mean just deleting the folder? does that break stuff?
for comparison, my instance is only communities and no users (FAQ here: https://programming.dev/post/442419 )
it’s about 5 days old but not much activity
# du -h --max-depth=2 ./ 73M ./volumes/postgres 7.8M ./volumes/pictrs 8.0K ./volumes/lemmy-ui 81M ./volumes 81M ./
the latest backup is only 7.4 MB zipped, I use this backup script
#! /bin/bash # https://join-lemmy.org/docs/administration/backup_and_restore.html#a-sample-backup-script now=$(date +"%Y-%m-%d_%H.%M.%S") cd ~/lemmy && (docker-compose exec -T postgres pg_dumpall -c -U lemmy 1> dump.sql 2> dump.errors) cd ~/lemmy && zip -r9 ~/bak-lemmy-$now.zip ./ --exclude "volumes/postgres/*" rm -f ~/lemmy/dump.sql
like I said though, not much activity
and here’s how my backup zip files have grown in size over time
6.6M bak-lemmy-2023-07-08_03.00.01.zip 6.6M bak-lemmy-2023-07-08_06.00.01.zip 6.6M bak-lemmy-2023-07-08_09.00.01.zip 6.6M bak-lemmy-2023-07-08_12.00.01.zip 6.6M bak-lemmy-2023-07-08_15.00.01.zip 6.6M bak-lemmy-2023-07-08_18.00.01.zip 6.6M bak-lemmy-2023-07-08_21.00.01.zip 6.6M bak-lemmy-2023-07-09_00.00.02.zip 6.6M bak-lemmy-2023-07-09_03.00.01.zip 6.6M bak-lemmy-2023-07-09_06.00.01.zip 6.6M bak-lemmy-2023-07-09_09.00.01.zip 6.6M bak-lemmy-2023-07-09_12.00.01.zip 6.6M bak-lemmy-2023-07-09_15.00.01.zip 6.6M bak-lemmy-2023-07-09_18.00.01.zip 6.6M bak-lemmy-2023-07-09_21.00.01.zip 6.7M bak-lemmy-2023-07-10_00.00.01.zip 6.7M bak-lemmy-2023-07-10_03.00.01.zip 6.7M bak-lemmy-2023-07-10_06.00.01.zip 6.7M bak-lemmy-2023-07-10_09.00.01.zip 6.8M bak-lemmy-2023-07-10_12.00.01.zip 6.8M bak-lemmy-2023-07-10_15.00.01.zip 7.0M bak-lemmy-2023-07-10_18.00.01.zip 7.0M bak-lemmy-2023-07-10_21.00.01.zip 7.0M bak-lemmy-2023-07-11_00.00.01.zip 7.0M bak-lemmy-2023-07-11_03.00.01.zip 7.0M bak-lemmy-2023-07-11_06.00.01.zip 7.0M bak-lemmy-2023-07-11_09.00.01.zip 7.0M bak-lemmy-2023-07-11_12.00.01.zip 7.0M bak-lemmy-2023-07-11_15.00.01.zip 7.4M bak-lemmy-2023-07-11_18.00.01.zip
Tested deleting the pictrs/files/001 which was at that point at ~150mb. And yes, it broke stuff. For example my Avatar and a local test community logo. To “purge” it, something would need to compare it to the database i guess. My size probably comes from /c/memes 😅
I thought remote images aren’t cached localy 🤔
I think it might just be thumbnails that are cached locally? And obviously your own banner and local images
Looks like random? Both images are from meme@lemmy.ml didn’t post on any of these 2 threads. It would be nice to know why sometimes its get copied. https://lemmy.ml/pictrs/image/b2e8d1e4-b638-4d04-9710-9fa28756261d.jpeg?format=webp https://lemmy.gwa.app/pictrs/image/962b1734-d524-468c-b3f9-4d615f91f094.png?format=webp
The interesting thing is, both posts originated from .world and when opened from the original sources are also hosted the.
https://lemmy.world/post/1318708
https://lemmy.gwa.app/post/1437
https://lemmy.world/post/1348802
https://lemmy.gwa.app/post/2310