You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi @Yumin-Sun-00 could you maybe try running Ctrl + Shift + P or Command + Shift + P on MacOS and type Dev Containers: Rebuild and Reopen in Container?
Hi, I am not sure, if the problem originates from the memory config of your VM instance.
Can you give some more details about your setup. I understood RHEL8.
Which version of Docker do you use? Which version of VSCode?
If you use the way Jens has described before, there also is an option "Dev Containers: Show container log"
Can you please provide that? This would probably give us more background.
Or was the log snippet above the entire log?
If the rebuild step from above did not work, you may want to try to delete the stopped container instance and image directly via docker commands and completely start from scratch.
Hi, I am not sure, if the problem originates from the memory config of your VM instance.
Can you give some more details about your setup. I understood RHEL8. Which version of Docker do you use? Which version of VSCode?
If you use the way Jens has described before, there also is an option "Dev Containers: Show container log" Can you please provide that? This would probably give us more background. Or was the log snippet above the entire log?
If the rebuild step from above did not work, you may want to try to delete the stopped container instance and image directly via docker commands and completely start from scratch.
# docker --version
Docker version 27.3.1, build ce12230
VS Code version: 1.95.2 (user setup)
Dev Containers version: v0.388.0
Delete images and containers I have to keep everything clean before rebuild:
# docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
hello-world latest d2c94e258dcb 18 months ago 13.3kB
# docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
b639d9e02e9c hello-world "/hello" About an hour ago Exited (0) About an hour ago amazing_torvalds
Now: In VS Code, reopen in containers --> exit code 137. Here is the logs:
Hi, I am not sure, if the problem originates from the memory config of your VM instance.
Can you give some more details about your setup. I understood RHEL8. Which version of Docker do you use? Which version of VSCode?
If you use the way Jens has described before, there also is an option "Dev Containers: Show container log" Can you please provide that? This would probably give us more background. Or was the log snippet above the entire log?
If the rebuild step from above did not work, you may want to try to delete the stopped container instance and image directly via docker commands and completely start from scratch.
In general I am not sure if it makes a lot of sense to deploy a VM instance only for this purpose. An complete alternative approach would be to use GitHub Codespaces for doing the lab. Here is a link how to set one up: https://github.com/codespaces/new?skip_quickstart=true&machine=basicLinux32gb&repo=751957128&ref=main&devcontainer_path=.devcontainer%2Fdevcontainer.json
Here is some general codespace documentation, if this is new for you: https://docs.github.com/en/codespaces
# docker --version
Docker version 27.3.1, build ce12230
VS Code version: 1.95.2 (user setup)
Dev Containers version: v0.388.0
Delete images and containers I have to keep everything clean before rebuild:
# docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
hello-world latest d2c94e258dcb 18 months ago 13.3kB
# docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
b639d9e02e9c hello-world "/hello" About an hour ago Exited (0) About an hour ago amazing_torvalds
Now: In VS Code, reopen in containers --> exit code 137. Here is the logs:
You could try to set the save directory in a startup command like this postStartCommand": "git config --global --add safe.directory ${containerWorkspaceFolder} in devcontainer.json.
thanks for the updates here. We were still unable to reproduce this behaviour on our end.
If we get the chance, can you let us some details, about the VM and in particular the OS and settings.
I would really like to figure this out before more people step into these problems.
I was also thinking about it is because of VM workstation, therefore at the beginning I provision a ec2 instance to try again.
I think this is more related to Git new security features, that the owner of files must match to the user who perform git operations. I found in github there were/are issues/discussions about this new features, some solutions are proposed. But it seems like there are still people suffering from this.
It sounds like it is about when .gitconfig will be copy pasted from host into the container.
From this perspective, I think it is less relevant to VMs, though at the beginning I indeed had such idea, therefore you see I provisioned a new ec2 instance.
I see 2 options to fix this, however I am not familiar with docker container.
What could make sure .gitconfig will be copied from local host into container.
Create a user inside container that matches the owner of repository.
I just roughly say out my opinions, but I do not know whether this is promising.
BTW the codes pace things also not work by me, not sure whether it is because I am using my company laptop....
You could also maybe try to download the repo as a zip file or to just delete the .git/ folder
Hey, that is a good idea! I get rid of the exit code 128 now 🎶🤩🥳
However, it still not work 🥲😂😂😂😂😂😂😂😂 Here is the log...I am done with this, I have to move on. It is a nice tutorial though that I wish to have a look very long time.
Activity
[-]Docker build container: Exit code 137[/-][+]DevContainer: Exit code 137[/+]Yumin-Sun-00 commentedon Nov 9, 2024
After adding these runArgs
in devcontainer.json the error messages changes to
JenSeReal commentedon Nov 10, 2024
Hi @Yumin-Sun-00 could you maybe try running Ctrl + Shift + P or Command + Shift + P on MacOS and type
Dev Containers: Rebuild and Reopen in Container?Yumin-Sun-00 commentedon Nov 10, 2024
Hi @JenSeReal
I provision a EC2 instance with Instance type
m4.2xlarge, which meansAnd I remove my changes and have tried:
Dev Containers: Rebuild and Reopen in Container
Dev Containers: Rebuild Without Cache and Reopen in Container
Both returns with
exit code 137.Yumin-Sun-00 commentedon Nov 10, 2024
Maybe more system informarion of my EC2 instance might helpful?
maeddes commentedon Nov 10, 2024
Hi, I am not sure, if the problem originates from the memory config of your VM instance.
Can you give some more details about your setup. I understood RHEL8.
Which version of Docker do you use? Which version of VSCode?
If you use the way Jens has described before, there also is an option "Dev Containers: Show container log"
Can you please provide that? This would probably give us more background.
Or was the log snippet above the entire log?
If the rebuild step from above did not work, you may want to try to delete the stopped container instance and image directly via docker commands and completely start from scratch.
In general I am not sure if it makes a lot of sense to deploy a VM instance only for this purpose.
An complete alternative approach would be to use GitHub Codespaces for doing the lab. Here is a link how to set one up:
https://github.com/codespaces/new?skip_quickstart=true&machine=basicLinux32gb&repo=751957128&ref=main&devcontainer_path=.devcontainer%2Fdevcontainer.json
Here is some general codespace documentation, if this is new for you: https://docs.github.com/en/codespaces
Yumin-Sun-00 commentedon Nov 10, 2024
Hi @maeddes ,
VS Code version: 1.95.2 (user setup)
Dev Containers version: v0.388.0
Delete images and containers I have to keep everything clean before rebuild:
Now: In VS Code, reopen in containers --> exit code 137. Here is the logs:
[1180 ms] Dev Containers 0.388.0 ov.txt
Thanks for providing alternative. I will have a look now.
Yumin-Sun-00 commentedon Nov 10, 2024
I found in the logs there is another exit code 128 before exit code 137:
fatal: detected dubious ownership in repository at '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/workspaces/LFS148-code'
To add an exception for this directory, call:
Do you think this could be the cause?
maeddes commentedon Nov 10, 2024
I think it could be, but I can't verify (yet). We've never seen this issue and we can't reproduce it
JenSeReal commentedon Nov 11, 2024
You could try to set the save directory in a startup command like this
postStartCommand": "git config --global --add safe.directory ${containerWorkspaceFolder}indevcontainer.json.Yumin-Sun-00 commentedon Nov 12, 2024
Hi @JenSeReal ,
I have tried different versions of such commands
None of them worked. I currently give up on this, and will try the codespaces as maeddes recommended.
Thanks for your support, I am appreciated on that.
maeddes commentedon Nov 12, 2024
Hi @Yumin-Sun-00
thanks for the updates here. We were still unable to reproduce this behaviour on our end.
If we get the chance, can you let us some details, about the VM and in particular the OS and settings.
I would really like to figure this out before more people step into these problems.
many thanks for patience here!
Yumin-Sun-00 commentedon Nov 12, 2024
Hi @maeddes
I was also thinking about it is because of VM workstation, therefore at the beginning I provision a ec2 instance to try again.
I think this is more related to Git new security features, that the owner of files must match to the user who perform git operations. I found in github there were/are issues/discussions about this new features, some solutions are proposed. But it seems like there are still people suffering from this.
It sounds like it is about when .gitconfig will be copy pasted from host into the container.
From this perspective, I think it is less relevant to VMs, though at the beginning I indeed had such idea, therefore you see I provisioned a new ec2 instance.
I see 2 options to fix this, however I am not familiar with docker container.
I just roughly say out my opinions, but I do not know whether this is promising.
BTW the codes pace things also not work by me, not sure whether it is because I am using my company laptop....
Have a good day!
🍻
JenSeReal commentedon Nov 12, 2024
You could also maybe try to download the repo as a zip file or to just delete the
.git/folderYumin-Sun-00 commentedon Nov 12, 2024
Hey, that is a good idea! I get rid of the exit code 128 now 🎶🤩🥳
However, it still not work 🥲😂😂😂😂😂😂😂😂 Here is the log...I am done with this, I have to move on. It is a nice tutorial though that I wish to have a look very long time.
[2024-11-12T152355.546Z] Dev Contai.txt
maeddes commentedon Nov 12, 2024
first of all thanks for providing all the feedback here. If we can resolve it we will reach out to you. Sorry we couldn't make it work yet
Nazario-DApote commentedon Feb 22, 2026
Hi,
I had the same problem on my Fedora 43. I solved it with these changes:
devcontainer.json
entrypoint.sh
i don't know how cannot attach the dockfile so i paste the content here:
FROM maven:3-eclipse-temurin-21
ARG DEBIAN_FRONTEND=noninteractive
ENV PIP_ROOT_USER_ACTION=ignore
RUN set -eux;
apt-get update;
apt-get install -y --no-install-recommends
ca-certificates
curl
bash
git
build-essential
python3.12
python-is-python3
python3-dev
python3-pip
python3-setuptools
python3-wheel
python3-psutil
docker.io;
rm -rf /var/lib/apt/lists/*
ENV JAVA_HOME=/usr/lib/jvm/temurin-21-jdk-amd64
ENV PATH="$JAVA_HOME/bin:$PATH"
RUN rm -f /usr/lib/python3*/EXTERNALLY-MANAGED
COPY requirements.txt /tmp/requirements.txt
RUN pip install --no-cache-dir -r /tmp/requirements.txt &&
rm /tmp/requirements.txt
WORKDIR /workspace
EXPOSE 5000
CMD ["bash"]