docker-compose exec container error #11081
Comments
|
Can you provide more details on your environment - how is Podman being run (root or rootless, from a systemd service or directly, installed from a package or built from source?). Can you also do an |
|
Fedora 34 vanilla install; podman rootless via systemd as a user service; regular 0755 root:root flags. |
|
Can you run containers with the Podman CLI ( |
|
The stack was started with podman, yes. I just cannot connect using docker-compose exec to the running containers. You see in the debug output that podman is indeed running containers. |
|
Did you try running containers without using compose? Can you try a |
|
Yes, and that works. I can also podman exec into a container started by compose, I just cannot exec from docker-compose. |
|
OK. I can't replicate this locally, so this is probably environment-related. Can you start the Podman service manually with debug-level logging enab;ed ( |
|
What about cgroup kernel settings? |
|
I doubt this is cgroup related, given this is an EPERM trying to run a new
process.
…On Fri, Aug 6, 2021 at 03:35 Henrik Holst ***@***.***> wrote:
What about cgroup kernel settings?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#11081 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AB3AOCE6EF4JPIEYLSDU4LDT3OGENANCNFSM5BHHRZ7Q>
.
Triage notifications on the go with GitHub Mobile for iOS
<https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675>
or Android
<https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email>
.
|
|
A friendly reminder that this issue had no activity for 30 days. |
|
@hholst80 Did you figure out any additional information on this? |
|
Friendly ping |
|
"Did you find out any additional information about this". No, I shelved the podman project for now as Docker engine has working support for this |
|
Thanks, @hholst80! I am closing this issue for now since we'd need more data. Feel free to reopen or comment in the future. |
|
I also run the same issue. Seems like $ podman-compose --help
usage: podman-compose [-h] [-f FILE] [-p PROJECT_NAME]
[--podman-path PODMAN_PATH] [--no-ansi] [--no-cleanup]
[--dry-run]
[-t {1pod,1podfw,hostnet,cntnet,publishall,identity}]
{pull,push,build,up,down,run,start,stop,restart} ...
optional arguments:
-h, --help show this help message and exit
-f FILE, --file FILE Specify an alternate compose file (default: docker-
compose.yml)
-p PROJECT_NAME, --project-name PROJECT_NAME
Specify an alternate project name (default: directory
name)
--podman-path PODMAN_PATH
Specify an alternate path to podman (default: use
location in $PATH variable)
--no-ansi Do not print ANSI control characters
--no-cleanup Do not stop and remove existing pod & containers
--dry-run No action; perform a simulation of commands
-t {1pod,1podfw,hostnet,cntnet,publishall,identity}, --transform_policy {1pod,1podfw,hostnet,cntnet,publishall,identity}
how to translate docker compose to podman
[1pod|hostnet|accurate]
command:
{pull,push,build,up,down,run,start,stop,restart}
pull pull stack images
push push stack images
build build stack images
up Create and start the entire stack or some of its
services
down tear down entire stack
run create a container similar to a service to run a one-
off command
start start specific services
stop stop specific services
restart restart specific services
$ podman --version
podman version 3.4.0related to containers/podman-compose#71 EDIT: I got it works by replacing podman-compose with docker-compose (without docker) then set the DOCKER_HOST to podman.sock.After removing podman-compose then install docker-compose. systemctl --user start podman.service
export DOCKER_HOST="unix://$XDG_RUNTIME_DIR/podman/podman.sock"
|
|
podman-compose is not part of Podman, it is a separate project. |
|
I have a very similar issue. Fresh install of Fedora 35 with both podman and docker-compose installed with dnf. I tried to follow the steps recommended by @mheon, and here are the logs being printed when I do Could this 👇 be the issue? |
|
Potentially? Is the error exactly the same as the one here ( Anything else in the logs? Dmesg? |
|
I'm not sure. I don't get an error message when running |
|
Log is empty (I assume you mean the output of |
|
@mheon Any tips on how I can further debug the issue? Or should I open a different issue to track this since I don't get the same error message? |
|
Full logs from a Podman API server ( |
|
@mheon here's a log for |
|
What Podman version are you on? I swear we fixed this (the cgroup parsing issue) recently. If not, I think we can zero in on that as the most likely culprit, nothing else jumps out - the network configuration issues are a bit worrying but are tied to other requests that seem to succeed. |
|
I'm on 3.4.4, installed from fedora's repositories. |
|
Alright, that's definitely new enough that it should have been fixed... I'm going to reopen this bug given all of this, we have a clear issue identified. |
|
@mheon couple of points here ... with docker-compose exec I now get a return code of 0 but also no output ... and now I only see the cgroups error when using podman exec. |
|
A friendly reminder that this issue had no activity for 30 days. |
|
A friendly reminder that this issue had no activity for 30 days. |
|
Don't believe so |
|
A friendly reminder that this issue had no activity for 30 days. |
|
@mheon was this one of the problems you fixed with docker-compose2? |
|
Nope, but I think we can close - we only have one report of the original issue and it's been almost a year. |
github-actions
bot
added
the
locked - please file new issue/PR
Description
docker-compose exec won't work fails with
Steps to reproduce the issue:
docker-compose up -d
docker-compose exec container sh
Describe the results you received:
Output of
podman version:Output of
podman info --debug:Package info (e.g. output of
rpm -q podmanorapt list podman):Minimal example docker-compose.yml
The text was updated successfully, but these errors were encountered: