Auto Download failing with HTTP/2 stream 1 was not closed cleanly: INTERNAL_ERROR

[mrgrosser]

Looking up latest version...
Downloading Bedrock server version 1.21.30.03 ...
curl: (92) HTTP/2 stream 1 was not closed cleanly: INTERNAL_ERROR (err 2)
ERROR failed to download from https://www.minecraft.net/bedrockdedicatedserver/bin-linux/bedrock-server-1.21.30.03.zip
Double check that the given VERSION is valid
DEBU[0000] Using /data to match uid and gid
DEBU[0000] Resolved UID=0 from match path
DEBU[0000] Resolved GID=0 from match path

WGET fails from ubuntu 22.04 server. Made sure to apt update and upgrade everything. Restarted. Manually tried to WGET,
wget https://www.minecraft.net/bedrockdedicatedserver/bin-linux/bedrock-server-1.21.30.03.zip
--2024-09-23 11:48:12-- https://www.minecraft.net/bedrockdedicatedserver/bin-linux/bedrock-server-1.21.30.03.zip
Resolving www.minecraft.net (www.minecraft.net)... 23.76.205.33, 23.76.205.26
Connecting to www.minecraft.net (www.minecraft.net)|23.76.205.33|:443... connected.
HTTP request sent, awaiting response...
Tried putting that link in my browser and it downloads fine. Wondering if Minecraft/Microsoft is blocking that agent id now and it must come from a browser? Is there a way to just manually download and put it somewhere in the server for now to bypass?

[jacokok]

I honestly think you are onto something, without user agent it does not work and works fine if any user agent is available.

[jacokok]

I created simple pull request to include agent: #459
Tested locally and change seems to fix this problem.

[mrgrosser]

That fixed it! Thanks for the merge. Server is up and running.

[josephdpurcell]

Ay carumba! Thanks for figuring this out! Here was my test:

I modified my docker-compose.yml to have an entrypoint:

services:
  minecraft:
    entrypoint: ["bash"]

That allowed me to attach to the container because currently it fails and keeps restarting.

Then I did docker attach minecraft and confirmed that making the request failed:

root@af2456a50f08:/data# curl -I https://www.minecraft.net/bedrockdedicatedserver/bin-linux/bedrock-server-1.21.31.04.zip
curl: (92) HTTP/2 stream 1 was not closed cleanly: INTERNAL_ERROR (err 2)

Then I tested the -A flag like so:

root@af2456a50f08:/data# curl -A "itzg/minecraft-bedrock-server" -I https://www.minecraft.net/bedrockdedicatedserver/bin-linux/bedrock-server-1.21.31.04.zip
HTTP/2 200 
content-length: 56027291
content-type: application/x-zip-compressed
content-md5: gPPO/pbSYdvJnG61s5DuDA==
last-modified: Mon, 30 Sep 2024 15:45:49 GMT
accept-ranges: bytes
etag: "0x8DCE166F5659C92"
server: Windows-Azure-Web/1.0 Microsoft-HTTPAPI/2.0
x-ms-request-id: d400f614-501e-007b-7d51-138a03000000
x-ms-version: 2018-03-28
mpulse_cdn_cache: HIT
mpulse_origin_time: 0
cache-control: max-age=160696
expires: Mon, 07 Oct 2024 15:57:23 GMT
date: Sat, 05 Oct 2024 19:19:07 GMT
server-timing: ak_p; desc="1728155946938_1749913683_83833968_9026_15587_18_40_15";dur=1

@itzg THANK YOU SOOOO much for this project and providing a way to track these types of issues and for issuing new fixes so quickly. I'll happily patiently await a new image.

[itzg]

@josephdpurcell the fixed image had already been pushed.

[josephdpurcell]

Woohoo!

I'm not sure the best way to do this, I did docker compose up and I thought it usually runs the new container? But, it didn't, so I did docker compose pull and then docker compose up and it worked.

THANK YOU!!!

[itzg]

Yep, a pull is always needed. Your approach is exactly right. I believe Docker's intent is to avoid surprise updates.

[tuknet2018]

I have the same issue. I’ve tried using Podman and Docker on different servers, and while all the containers have internet access, the problem persists: curl cannot download the .zip file. I attempted to change the entry point to /bin/sh and manually download the .zip with the agent options for curl. If I try curl without the -A parameter, it doesn't work. Downloading the file normally via the endpoint seems impossible.

Create a new server, pull new images and the latest , and the same problem continue.

[root@podman world_seed]# podman images
REPOSITORY TAG IMAGE ID CREATED SIZE
docker.io/itzg/minecraft-bedrock-server 2024.9.0 507cb46531db 13 days ago 185 MB
docker.io/itzg/minecraft-bedrock-server latest 28e90626f0e4 2 weeks ago 185 MB

[root@docker world_seed]# docker images| grep bedrock
docker.io/itzg/minecraft-bedrock-server 2024.9.0 507cb46531db 13 days ago 180 MB
docker.io/itzg/minecraft-bedrock-server latest 28e90626f0e4 2 weeks ago 180 MB

On centos or redhat , after pull the image, a recreate the container the problem continue.

[root@podman world_seed]# podman run -d -it --name mc -e EULA=TRUE -e LEVEL_SEED=822815728642535760 -p 88888:19132/udp -v /world_seed:/data itzg/minecraft-bedrock-server:latest
7b17f4b032d92d6d9d92bb0e955482e144a93adefa9882db034ce67dabb32174
[root@podman world_seed]# podman logs mc
DEBU[0000] Using /data to match uid and gid
DEBU[0000] Resolved UID=65534 from match path
DEBU[0000] Resolved GID=65534 from match path