-
Notifications
You must be signed in to change notification settings - Fork 3
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Terminating a Docker container with 1 process takes > 10 sec #8
Comments
Experiencing similar issue, except sometimes terminating gets stuck for over 4 minutes. I had to shell into the pod and manually 'kill -9' the socklog process. |
I took a quick look at the code and I think I know what is happening. The correct fix is to rewrite socklog as an asynchronous event loop, with "a message arrived on the socket" and "a signal arrived" (or "the selfpipe is readable" in skalibs idiom) as the two processable events. Well I guess it's a good incentive for me to finally add a syslog reader to s6, but in the meantime, @jprjr can you please fix this? |
Done - just pushed up socklog v2.2.3 with the quick and dirty fix, and socklog-overlay v3.1.2 with the updated socklog binary. |
I recently started seeing this issue crop up. I'm not entirely sure why though because I see this hasn't been updated in months, while my issue started occuring 3 days ago. My CI/CD pipeline builds a container and tests the shutdown mechanism daily which has stopped working suddenly with no changes to any of the service files. Entering into the container, I see all the processes have been reaped except for socklog. Forcibly killing it, immediately allows the container to terminate. |
The latest version of s6, which was released yesterday, includes a |
Got the latest binaries available at https://github.com/just-containers/skaware/releases/tag/v2.0.7 It'll be a bit before I can get down to testing/verifying everything, I'm heading out of town tomorrow. My plan is to keep this repo and just remove binaries from the released tarballs, to keep the install process the same. I considered merging it into I also need to sit down and finally address just-containers/s6-overlay#329 - when I get back I'll get that done, and bump the major version for both projects. |
I'm frustrated at the time it takes my container to shut down when using the S6 socklog overlay service. When i use the socklog service it takes ~10sec but when i don't it takes ~700ms.
** I realize in the example below nginx does not need syslog. The real service i want to use, postfix, does use syslog so that is why i am using socklog. I used the example below because that is what s6 uses for their example.
Dockerfile
Build the image:
docker build -t test .
Run the Image:
docker run --name test --rm -d test
Stop the Image:
time docker stop test
The text was updated successfully, but these errors were encountered: