imputnet
Can't download stuff from YouTube, cobalt says it's disabled or returns 0 byte files
bug description
every time i download any video the bytes are 0 off youtube
reproduction steps
1 go to youtube 2 copy and paste link into cobalt 3 hit the arrows to download 4 see that the files have 0 bytes on them
screenshots
links
No response
platform information
window 10 on chrome/google on ios
additional context
No response
we've been aware of the issue since the very beginning (approx 4 hours ago). it's not caused by us and is not an issue with cobalt's code, youtube blocked all of our ip ranges this morning.
we're currently doing our best trying to come up with a solution to restore downloads from youtube asap. we're sorry for any inconveniences caused in the meantime :(
as a temporary measure, you can try self-hosting a cobalt processing instance on your computer or home server: https://github.com/imputnet/cobalt/blob/main/docs/run-an-instance.md
i've temporarily disabled youtube support on the main instance, so that you don't get 0 byte files.
beside ip blocking they send abuse messages, be careful
i'm soooo scared of a baseless threat by google (thank you for the warning though)
Might it be possible to setup Mullvad on the processing servers to bypass the IP blocks? I know a VPN is a bit tricky on servers due to incoming traffic, but not impossible.
@Sainan I actually have Gluetun setup in a docker container which uses Mullvad, I then connect it to the Cobalt API instance that is also running in a docker container. This worked for me but not for very long, the IP's eventually get blocked (pretty quickily too).
Set it up to automatically switch Mullvad servers every hour. 🙈
I assume you're joking, but if not, rotating proxies might actually be a better solution than a VPN.
I'm not joking. It's what we did for checking a certain anti-cheat provider for information about bans. They did not like that we were providing transparency into their business. After deploying this Mullvad approach, they seemed to stop bothering.
Wouldn't you eventually run out of servers to use on Mullvad? Plus, YouTube could end up blocking their ASN altogether. A rotating proxy might be a more sustainable solution for something like this.
The idea is that legit users still use Mullvad, so blocking them outright might be less ideal.
I think you're right in that they wouldn't block them outright. Still a possiblity since it is Google we're talking about. Regardless I think running a VPN over the entire instance would be slow, potentially cause issues with other services and be somewhat useless since the time it takes YouTube to block the IP is inconsistent, so changing server every 1 hour wouldn't be a solution.
The idea is that legit users still use Mullvad, so blocking them outright might be less ideal.
true, they also move their servers around quite often
as others have mentioned, it's been fixed for a few days (since evening of 12th), but we made sure to get backups so this doesn't happen again in the future. can never be sure with youtube though.
still getting 0b over here. ig youtube links are still broken
anyways, ight imma go selfhost cobalt
Hello! i created an account just to say that issue is still persistent with YouTube links. I was able to download SoundCloud links but YouTube doesn't seem to be working at the moment again.
Wouldn't you eventually run out of servers to use on Mullvad? Plus, YouTube could end up blocking their ASN altogether. A rotating proxy might be a more sustainable solution for something like this.
@wukko Did you deploy this solution? Can you please share with us how you fixed it? Thank you so much!
I appear to be experiencing an issue like this, I am getting a 0 byte file, but most peculiarly, what I am being provided is a 0 byte .txt file, for whatever reason.
Now, I want to be clear also that I have not somehow managed to change anything on my end to make my computer think that .mp3 files are to be opened with Notepad.exe, because that is not what is happening on my end. It seems that either Cobalt or Firefox interprets the file in question as being a .txt file.
I appear to be experiencing an issue like this
@Chonkblonk can you try again?
