"deployment failed" but tasks are updated about ecs-deploy HOT 8 CLOSED

Hi @jamesdixon, did you read this section already?
https://github.com/fabfuel/ecs-deploy#troubleshooting

Best
Fabian

from ecs-deploy.

Did you edit the output? Which version of eco-deploy do you use?

from ecs-deploy.

@fabfuel we're running the latest version of ecs-deploy. I'm not sure what you mean by "did you edit the output?" Can you explain?

Yes, we're already using a timeout -- this is failing very quickly after the "Deploying new task definition" message appears.

Thank you!

from ecs-deploy.

Hi,

I meant, if this is the full output of the script or if you edited or removed something – because it looks unfamiliar.

Do you have any Events during the deployment in the Events tab of the service in the AWS ECS console?

Best
Fabian

from ecs-deploy.

"Very quickly" can't be, because every dot after "Deploying new task definition" means, it was waiting for one second. This took at least 26 seconds until it failed.

But what I'm missing is the error itself. Do you see this error message?

Deployment failed due to timeout. Please see: https://github.com/fabfuel/ecs-deploy#timeout

What command do you use for triggering the deployment?

from ecs-deploy.

@fabfuel
We have similar situation, but in our case increasing timeout won't help, because we deploy image which is expected to fail (we have implemented health-checks in our containers - so the new task definition should fail end exit with non-zero status, and the previous one should live all the time). At this point, after failure new task definition should be deregistered - but is not.
What do you thing about new flag (eg. deregister-new-task-on-failure)?
We could provide PR with such feature.

from ecs-deploy.

@fabfuel thanks for the tip on the "Events" section. Missed that!

Turns out the instance didn't have enough memory given the memory limit on the container.

Sorry for the false alarm!

from ecs-deploy.

Hi @gadomsky, have you tried the --rollback option?

from ecs-deploy.