Why is docker-compose failing with ERROR internal load metadata suddenly?

I've been running docker-compose build for days, many times per day, and haven't changed my DOCKERFILEs or docker-compose.yml. Suddenly an hour ago I started getting this:

Building frontdesk-api failed to get console mode for stdout: The handle is invalid. [+] Building 10.0s (3/4) => [internal] load build definition from Dockerfile 0.0s [+] Building 10.1s (4/4) FINISHED => [internal] load build definition from Dockerfile 0.0s => => transferring dockerfile: 32B 0.0s => [internal] load .dockerignore 0.0s => => transferring context: 2B 0.0s => ERROR [internal] load metadata for 10.0sailed to do request: Head dial tcp: lookup mcr.microsoft.com on 192.168.65.5:53: => [internal] load metadata for 0.0s ------ > [internal] load metadata for ------ ERROR: Service 'frontdesk-api' failed to build

Things I've tried:

Running it again
docker rm -f $(docker ps -a -q)
docker login
Different SDK image

Here is the DOCKERFILE:

FROM AS base WORKDIR /app EXPOSE 80 EXPOSE 443 FROM AS build WORKDIR /app # run this from repository root COPY ./ ./ #RUN ls -lha . RUN echo 'Building FrontDesk container' WORKDIR /app/FrontDesk/src/FrontDesk.Api #RUN ls -lha . RUN dotnet restore RUN dotnet build "FrontDesk.Api.csproj" -c Release -o /app/build FROM build AS publish RUN dotnet publish "FrontDesk.Api.csproj" -c Release -o /app/publish FROM base AS final WORKDIR /app COPY --from=publish /app/publish . ENTRYPOINT ["dotnet", "FrontDesk.Api.dll"]

How can I get my build working again?

33 Answers

1 2 Next

the solution that worked for me is to delete the .docker/config.json by running

rm ~/.docker/config.json

then docker-compose up your-services should work

For linux

Step 1: sudo vi ~/.docker/config.json

You will see something like this:

 { "credsStore": "desktop.exe" }

Step 2: Update credsStore to credStore

Step 3: Run your services docker-compose up your-services

Windows:

Step 1: Go to C:\Users\Name\.docker

Step 2: Open config.json

Step 3: Update credsStore to credStore

Step 4: Run your services docker-compose up your-services

In my case, it was corporate VPN that caused this issue. Once I disabled the VPN, it worked fine. Not really a solution, I admit. You could disable the VPN for the duration of this command run and enable it back.

mcr.microsoft.com is down at the moment

I'm receiving several different errors when pulling:

% docker pull Error response from daemon: Get Service Unavailable % docker pull 5.0: Pulling from dotnet/sdk received unexpected HTTP status: 500 Internal Server Error

What worked for me on Mac OS X was editing the Docker Config:

sudo vi ~/.docker/config.json

Changing

"credsStore": "desktop"

"credStore": "desktop"

And running, in my case:

sudo docker build -t php-demo .

( to build the demo/sample app )

Could it be that, it's getting the wrong arch/platform to build the image? Try specifying one in the command:

docker build -t auth:1.0.0 --platform=linux/arm64 .

getting the current arch

arch

In my case the problem arose because I have Apple Silicon in my laptop, and there is no version of the image I want that is compiled for linux/arm64/v8.

To see if this is your issue, try docker pull <image> for more detailed error messages.

For me the problem was in the General settings.

uncheck Use Docker Compose V2 and click Apply & Restart

I was facing this issue while building docker container as:

Error:

 [internal] load metadata for ------ failed to solve with frontend dockerfile.v0: failed to create LLB definition: rpc error: code = Unknown desc = error getting credentials - err: exit status 1, out: `` Applications failed at ./run:build line 18.

Solution:

I went to Docker desktop-->settings-->Docker engine--> changed "buildkit": true to "buildkit": false

{ "builder": { "gc": { "defaultKeepStorage": "20GB", "enabled": true } }, "experimental": false, "features": { "buildkit": false } }

Now Apply and restart docker and build the application.

Once application build successfully, revert back the changes of docker engine "buildkit": false to "buildkit": true

{ "builder": { "gc": { "defaultKeepStorage": "20GB", "enabled": true } }, "experimental": false, "features": { "buildkit": true } }

and rebuild application

I think this has to do with issues trying to load the meta data of the image specified in the dockerfile from the dockerhub or the container repository of your choice. a quick fix is to check if your internet is up or you have necessary access. this is what resolve it for me.

Started getting this error out of nowhere, checked my WSL Integration and the toggle was off.

After enabling and restarting it again it started working.

I had this error pop up recently and I tried changing nameservers etc. Finally restarting docker helped. sudo systemctl restart docker

It was proxy settings for me. I had proxy enabled in docker desktop. Disabling worked for me.

Yeap, faced with the same issue. As guys said - looks like it is their side issue or DNS is feeling bad. Try to pull it manually using docker pull few times in a row, it may help.

Also, try to switch your internet provider from cable/wireless to gsm for example if you have the ability to do it, also working solution.

For me it was because I was using docker-credential-helpers in ~/.docker/config.json.

{ "auths": { "ghcr.io": {}, "": {}, "registry-intl.eu-central-1.aliyuncs.com": {} }, "credsStore": "pass" }

So I needed to unlock the keystore, for example with the following command for index.docker.io:

pass show docker-credential-helpers/aHR0cHM6Ly9pbmRleC5kb2NrZXIuaW8vdjEv/USERNAME

In my case it was damaged .dockerignore file. Just try to delete it/rename it/recreate it.

The full error I got was:

[+] Building 2.2s (7/16) => [internal] load build definition from Dockerfile 0.0s => => transferring dockerfile: 1.31kB 0.0s => [internal] load .dockerignore 0.0s => [internal] load metadata for 0.9s => [internal] load metadata for 2.0s => [angular-built 1/8] FROM 0.0s => => resolve 0.0s => [stage-1 1/3] FROM 0.0s => => resolve 0.0s => ERROR [internal] load build context 0.0s => => transferring context: 0.0s ------ > [internal] load build context: ------ failed to solve: rpc error: code = Internal desc = stream terminated by RST_STREAM with error code: PROTOCOL_ERROR

For me what worked is deleting the "credsStroe" from the config.json file

In my case I was using docker compose to build from a local dockerfile. Turned out I had a COPY action in my dockerfile that was using a --from=x referencing an earlier FROM where I forgot to add the AS x.

Docker-compose didn't show any of this though. It wasnt util I ran docker build that it showed me the detailed info as to why it failed.

Happened the same to me after adding custom nameservers.

 [network] generateResolvConf = false

I reverted that line on

/etc/wsl.conf

and after doing wsl.exe --shutdown on the powersell and restart again docker, it worked fine again.

[Specific solution] Only if building from a virtual environment. Try deactivating it, event if it is the base one. e.g. for conda:

conda deactivate

no disk space

I just got the exact same error and in my case the problem was that I didn't have enough disk space. I'm running docker in a VM with 32 GB virtual disk and it just ran out of all the docker images. I freed up some space and the problem was solved.

Check your proxy settings. For me, this got resolved after proxy settings change

git config --global http.proxy proxy-eu.<name>.com:8080

name should be replaced

For me the problem was that the the token to access the registry was outdated. So make sure that docker login your-registryis working well.

Reference:

if you write code for a node like this FROM node: alpine then must need to pull

docker pull node

hopefully, it was work

After several weird attempts that were failing,

I just deleted the .docker folder

location: C:\Users\Name.docker

and it worked .... for me

Your session expired.

You have to log in again at

I had the similar error when building for multiple platforms

docker buildx build --platform linux/amd64,linux/arm64 -t name:tag . -o type=image

I tried for single platform to update the base image (linux/amd64 then linux/arm64)

docker buildx build --platform linux/amd64 -t name:tag . -o type=image

After done updating, I just canceled to run for both platforms and it works. In my case, there might be a bug when fetching latest base image for multiple platforms.

clearing disk space by deleting images & restarting docker did the trick for me.

I fixed it as below for windows machine,

Go to your C://Users/YOUR_USER_FOLDER
Get inside .docker folder
Remove config.json file
Using command prompt/command line/terminal, I ran docker login cmd, entered username/password and login was successful.
Above command created config.json again with below contents,
```
{ "auths": { " }, "credsStore": "wincred" } 
```