BentiGorlich

joined 1 year ago
MODERATOR OF
[–] [email protected] 15 points 1 week ago

Just for the people interested in mbin, here are a few magazine/community pointers:

[–] [email protected] 7 points 1 week ago

Glad I couldbe oh any help 😇

[–] [email protected] 11 points 1 week ago (2 children)

Glad it works again :)

 

Soso. Friedrich Merz ist also Kanzlerkandidat. Jetzt muss er sich ja nur noch entscheiden, für welche Partei er antreten möchte. Übrigens: Morgen (20.9.24) ist Klimastreik.

#Merz #Kaenguru #Klimastreik #MarcUweKling

[–] [email protected] 18 points 1 month ago

I am part of the mbin team and I am really tired of hearing shittalking about us without any reason. Nobody has any reason to call us anything but passionate.

When melroy started the fork, really weird accusations were thrown around without being based on anything...

[–] [email protected] 2 points 1 month ago (1 children)

Weil der Artikel davon ausgeht dass man beide postgresql Versionen parrallel laufen lassen kann und keine Option anbietet falls es nicht funktioniert. Eigentlich hätte ich laut dem Artikel keinen Datenverlust erleiden dürfen

Monitor the process, and upon success, confidently remove the older version.

sudo pg_dropcluster 14 main

Das heißt, soweit ich das verstehe, dass eigentlich die Daten bei dem neuen Cluster eingespielt werden sollten, das ist aber nicht passiert. Vermutlich einfach weil das neue 16er Cluster bei mir nicht starten konnte. Der Artikel erwähnt davon aber nichts...

 

Da gehirneimer heute mit Serverproblemen zu kämpfen hatte habe ich mir gedacht ich spiele mal die neuste Version der Datenbank Software (PostgreSQL) auf (von v14 -> v16). Dabei ist mir leider ein Missgeschick unterlaufen und ich habe die aktuelle Fassung der Datenbank gelöscht. Ich habe jetzt das letzte Backup wiederhergestellt und alles scheint wieder zu funktionieren, allerdings fehlen uns jetzt erstmal einige Posts... Ich entschuldige mich aufrichtig dafür, sowas kann leider mal passieren, weswegen eine gute Backup Strategie enorm wichtig ist...

Die Story

Heute morgen um 4 Uhr hat der Server angefangen mir eine zu hohe Auslastung zu melden. Ich habe mir erstmal nichts dabei gedacht, kann ja sein dass einfach einiges los war. Ich habe es erst gegen 9:30 Uhr untersucht. Da ist mir dann aufgefallen dass sehr viel CPU Zeit dafür drauf ging auf die Festplatte zu warten (IO-Wait), das hab ich so noch nicht gesehen und hat mich sehr verwundert. Ich habe die Caches geleert, den Server neu gestartet und mir nochmal die letzten Änderungen an der Konfiguration angesehen, aber mir ist nichts aufgefallen. Ich habe dennoch die Konfiguration für den Redis Cache zurück gedreht um Fehler auszuschließen, das hat jedoch nichts gebracht...

Nach einer Weile musste ich mich damit zufrieden geben die Hintergrunddienste zurück zu schrauben, damit der Server nicht komplett lahmgelegt wird, denn ich musste frühstücken und hatte etwas vor. Hetzner habe ich aber noch kontaktiert befor ich losgegangen bin.

Als ich dann später wieder Zeit hatte hab ich dann angefangen nach Lösungen zu suchen. Ich habe mir gedacht dass es helfen könnte die Datenbank zu aktualisieren, da dies das Hauptproblem war/ist. Die DB verursacht nunmal die meiste Arbeit für Festplatten. Die Seite die ich zurate gezogen habe hat ausdrücklich gesagt dass eine neue Version erstellt wird die die Daten der letzten Version weiter benutzt. Ich konnte die beiden Versionen nicht parallel laufen lassen (was eigentlich vorgesehen ist, damit man keinerlei downtime hat), da der Arbeitspeicher den der Server hat dafür einfach nicht ausreicht. Also hab ich probiert die alte Version zu stoppen damit ich die neue starten kann (da RAM frei wird). Der Befehl klang schon verdächtig danach dass die Version damit nicht nur heruntergefahren, sondern einfach gelöscht wird (der Befehl lautete pg_dropcluster 16 main --stop), aber ich habe mir gedacht "ist ja nicht schlimm, die neue Version nutzt ja einfach dieselben Daten". Tja denkste... die Daten wurden durch diesen Befehl einfach gelöscht.

Und so begann der Prozess der Wiederherstellung der DB der leider auch einige Stunden gedauert hat...

Der Artikel dazu den ich absolut nicht empfehlen kann: https://medium.com/@malymohsem/say-goodbye-to-downtime-upgrade-postgresql-14-to-16-effortlessly-with-pg-upgrade-42ef4dbf8524

[–] [email protected] 10 points 1 month ago (1 children)

AP has a tool for that called inbox forwarding and mastodon uses it for sharing the comments under posts. It works like this: you send a reply to a user with their follower collection as the recipient. You of course cannot know who is following that user, however they than just forward this reply to the follower collection, because the server knows that it has authority over that collection. https://www.w3.org/TR/activitypub/#inbox-forwarding

[–] [email protected] 39 points 1 month ago (3 children)

Its not really a "not playing well" with each other, it is just the mastodon works. That is the reason why most toots in my mastodon timeline have 0 favourites (upvotes) and only a few boosts... I don't know why they do it, because at the minimum the followers of that user should be notified about that like...

[–] [email protected] 4 points 1 month ago (1 children)

I promise that I will never just suddenly vanish unless I actually die 😅

[–] [email protected] 4 points 1 month ago (1 children)

I actually have a LiberaPay, Patreon and Kofi page and only 3 people have donated so far. And those accounts are also used to take donations for my 2 mbin servers and my mid sized mastodon server (>100 "real" MAU)...

[–] [email protected] 9 points 1 month ago

Your already helping a lot by helping other admins ❤️

[–] [email protected] 12 points 1 month ago (3 children)

I think we are getting there. Currently the biggest problem I see is that melroy and I are currently the main contributors with other experienced contributors gone or at least not regularly active (which is of course their right, but its just how it is). I do have a whole lot of ideas and plans for features and I am working on a lot simultaneously, though I wouldn't say that it is a burden, more like a fun hobby. Admittedly this hobby takes a lot of time though 😅

I am very happy that new contributors are currently dipping their toe in the water and I hope they do stick around. Also a lot of the server admins are active in the matrix chat to help others out. So in this regard we are not alone.

I was honestly shocked that debounced left so suddenly and basically completely vanished... I didn't think that anyone would do that... So honestly we cannot promise anything, but I think melroy will stay around and I certainly will as well 😇

[–] [email protected] 2 points 1 month ago

I had the exact same problem and the solution was to ask my ISP who then either just gave me a public IP (Vodafone) or asked for money so my network could be reached from the outside (Primerocom). So check whether there is an option with you ISP to get a "public" IP.

 

Just listened to this banger, love imminence ❤️

 

This is just a test thread :)

 

The biggest surprise for me was the https://hexbear.net count, an instance I hardly interact with.

beehaw.org hexbear.net lemdro.id lemmy.blahaj.zone lemmy.dbzer0.com lemmy.ml lemmy.one lemmy.world lemmynsfw.com sh.itjust.works sopuli.xyz
Community Count 6 33 1 1 1 14 1 39 2 1 1
Community Subscriber Count 133450 663204 17052 15907 53006 356460 16257 851950 33586 16006 14093

The data this is based on comes from https://lemmyverse.net where you can just download a full json of the data they have (I excluded all communities marked as "suspicious")

EDIT: The data if you sort by active users last month:

awful.systems feddit.org feddit.uk hexbear.net lemdro.id lemm.ee lemmy.blahaj.zone lemmy.ca lemmy.dbzer0.com lemmy.ml lemmy.world lemmy.wtf lemmy.zip mander.xyz sh.itjust.works slrpnk.net sopuli.xyz ttrpg.network
Community Count 1 2 2 1 1 3 1 3 3 10 57 1 3 1 5 3 2 1
Community Active Month Count 2616 7363 5289 2952 2898 8898 11422 14910 13752 54949 338384 3602 12020 11469 37365 10897 10070 4107

Community Count:

Community Users:

 
 

Hello everyone, I noticed that a lot of my music does not get any album art even though they correctly identified with MusicBrainz and the albums do have a cover on them there. Jellyfin does not have write permissions in the data directly where the music is, only read permissions. Though I would have expected that jellyfin uses the metadata folder for cover images not already present there. The reason Jellyfin doesn't have write permissions is that the folder is managed by nextcloud and that creates files with 750 so only read access for the group and jellyfin uses the group Does anyone know how I could solve the issue?

Log when trying to identify an album which should have edited the image:

[WRN] Unable to delete "/{{REDACTED}}/files/Musik/Everglow/Last Melody [Single]/Folder.jpg" System.UnauthorizedAccessException: Access to the path '/{{REDACTED}}/Musik/Everglow/Last Melody [Single]/Folder.jpg' is denied. ---> System.IO.IOException: Permission denied


End of inner exception stack trace


at System.IO.FileSystem.DeleteFile(String fullPath) at MediaBrowser.Providers.Manager.ItemImageProvider.PruneImages(BaseItem item, IReadOnlyList`1 images)

I tried giving Jellyfin write permissions on the folder and when identifying an album and it just decided to delete all tracks...

from the log after giving write permissions:

[INF] Removing item, Type: "Audio", Name: "RATATATA", Path: "/{{REDACTED}}/Musik/Babymetal/RATATATA/1 - RATATATA.mp3", Id: 72e62d6b-7afd-842e-8a72-b9142c53cdaf [INF] Setting provider id's to item ae9bfd0c-30e7-8943-c552-5f89475b4116-"RATATATA": [("MusicBrainzAlbum": "4a762493-80fe-4a52-b11a-ad6ecfcc8a59"), ("MusicBrainzReleaseGroup": "d61d121b-8952-40de-bef9-208b9ea870ec")] [INF] Setting provider id's to item ae9bfd0c-30e7-8943-c552-5f89475b4116-"RATATATA": [("MusicBrainzAlbum": "4a762493-80fe-4a52-b11a-ad6ecfcc8a59"), ("MusicBrainzReleaseGroup": "d61d121b-8952-40de-bef9-208b9ea870ec")] [INF] Removing item, Type: "Audio", Name: "STΦMP! (STΦMP!)", Path: "/{{REDACTED}}/Musik/Dreamcatcher/[VirtuouS]/3 - STΦMP! (STΦMP!).mp3", Id: 32dfa172-442f-af4d-9d6e-0ad4d9a736c6 [INF] Removing item, Type: "Audio", Name: "Intro : 7' Dreamcatcher (Intro : 7' Dreamcatcher)", Path: "/{{REDACTED}}/Musik/Dreamcatcher/[VirtuouS]/1 - Intro : 7' Dreamcatcher (Intro : 7' Dreamcatcher).mp3", Id: e83079d4-9c51-7f8b-1b55-c54d1df3787b [INF] Removing item, Type: "Audio", Name: "JUSTICE (JUSTICE)", Path: "/{{REDACTED}}/Musik/Dreamcatcher/[VirtuouS]/2 - JUSTICE (JUSTICE).mp3", Id: 8d977ec0-6f53-cbd9-65d2-45ddcc1b78d0 [INF] Removing item, Type: "Audio", Name: "2 Rings (2 Rings)", Path: "/{{REDACTED}}/Musik/Dreamcatcher/[VirtuouS]/4 - 2 Rings (2 Rings).mp3", Id: 987a8bdc-8c73-1a74-16ab-341dacf035c3 [INF] Removing item, Type: "Audio", Name: "Fireflies (Fireflies)", Path: "/{{REDACTED}}/Musik/Dreamcatcher/[VirtuouS]/5 - Fireflies (Fireflies).mp3", Id: fe1a2125-ad68-3b2f-fddf-7f3cabf43ccd

Any help is greately appreciated :)

17
submitted 2 months ago* (last edited 2 months ago) by [email protected] to c/[email protected]
 

I am coming from a Unity background and there I just had a component of some custom class in the scene which I could then easily get by calling FindInScene<CustomComponent> or something like that. Not in Godot this doesn't work, because I didn't find a way to get the actual class of an attached script. I always just get GDScript as the class name even though I did specify a custom one.

The information I want to save are things like: where to spawn players, how many laps will this race have, maybe save references to the spawned players, etc.

So how would I save this "meta" information to get by another script in Godot?

EDIT: here is an example: I have a main scene which can load different levels. When loading a level I need to get some information about that level, like: the available spawn points. Inside each level I have a node with a script attached to it that defined class_name LevelMeta and holds the meta information that I need when loading the level. How do I detect this script and the associated meta information?

 

I recently tried to implement push notifications, but it does not seem like fennec is supporting them. The same code works fine on vanilla firefox on mobile, Librewolf and Firefox on desktop.

Does anyone know whether Fennec should support push notifications?

 

Die Australier schlafen jetzt und die Briten haben unsere Flagge fast komplett ausgelöscht :O

 

Hallo zusammen, ich hatte ne Abmachung mit den Australiern dass wir die Flaggen vereinen. Ich find die Idee nach wie vor super und wollte euch das auch mitgeteilt haben :D

 

I've been thinking of getting that game, though I am not sure on the development pace and state it is currently in.

Generally I love city builders of any kind and the game looks super interesting to me (and it is currently on sale :D ). What are your thoughts?

view more: next ›