Closed justanr closed 7 years ago
Yep, this can happen depending on your situation. Can you give me more details about your process and your configuration? In particular, do you have any lifecycle checks for the running
state configured in Maestro?
Well now I feel like a putz, there is a lifecycle hook hidden in the very root template we're using that hits up an http endpoint, which isn't coming up within the allotted timeframe sometimes.
No worries. Glad you found where your problem came from!
We're receiving an exception from the StartTask that some containers failed to report a running status. However, when I look at the host daemon, the containers are running just fine.
I'm able to reproduce this by sleeping for 20 seconds in the entry point script, the actual trigger -- which is inconsistent at best -- is compiling django translations in the entrypoint (which isn't being done at build for reasons).
When I run locally with the sleep, docker ps it shows the container up, and running
docker inspect --format='{{.State}}' <id>
it shows running as the state.The error bubbles from maestro.play.tasks.StartTask and looks like the internal
_wait_for_status
helper is returning False; however, I'm not sure where the actual error is stemming from (docker-py, the container wrapper, etc).