Allow stop/kill to continue in the event that it cannot stop or kill one of the old containers #1356
Conversation
…one of the old containers. Also actually attempt to kill if stop has a bad exit code.
|
Thanks for the PR. 👍 for me. @josegonzalez thoughts here? |
|
Reminds me of that @akatrevorjay is there a way we can log a message when the kill fails? Maybe a hook call... |
|
Oh yeah. Where is that thing? Was it just in a gist somewhere or do you have it in a repo? The docker stop/kill loop gets |
|
Is there a way we can introduce general logging of this? |
|
We could trigger an event |
|
Lets do that then. I want to have something there for our users so that at least it can be logged/detected and taken care of, even if out of band. |
|
👍
|
|
@akatrevorjay let me know if you're comfortable adding the logging or if you'd prefer assistance |
|
@michaelshobbs pretty sure @akatrevorjay won't get to the logging bit, so if you want to take the reins and merge this, go ahead. |
In the event that we cannot stop a container, we should try to kill it, and if that then fails, we should continue to the next $oldid , otherwise one container not actually being there or not able to be stopped will prevent the rest from being cleaned up.
The way this is, if a container ever fails to stop, it will never kill it anyway, due to being executed with -e.
This fixes that as well with a simple or.
The bigger issue at hand here is that if you change your container process types around it doesn't really cleanup after itself, leaving CONTAINER.$type.$nid and PORT.$type.$nid files around, but I didn't get around to addressing that.