Open ookblah opened 5 years ago
Thanks for the bug report! Could you attach your deployfish.yml so we can try to replicate your issue here?
Sure thing, here's the relevant service from my deployfish.yml and output from the CLI. Don't mind the stuff about traefik. I'm basically spinning up a php service.
#PHP Service
- name: app
cluster: cv-staging
launch_type: FARGATE
execution_role: <service role arn>
network_mode: awsvpc
service_discovery:
namespace: cv-staging
name: app
dns_records:
type: A
ttl: 10
vpc_configuration:
subnets:
- <subnet 1>
- <subnet 2>
security_groups:
- <security group 1>
public_ip: ENABLED
cpu: 256
memory: 512
count: 2
family: app
containers:
- name: app
image: <php image>
labels:
- "traefik.enable=false"
logging:
driver: awslogs
options:
awslogs-group: cv-staging
awslogs-region: us-east-1
awslogs-stream-prefix: app
I'm able to initiate creating the service but then it repeatedly goes into this in my cli before failing deployment. Tasks in the ECS dashboard go into an endless cycle of activating, provisioning, etc. At first glance everything seems to run fine and the service appears to be up, but if you switch to the stopped tasks you see all the dead ones. If I remove the block about service discovery it works fine.
Deployment Desired Pending Running
PRIMARY 2 2 0
Service:
(service app) has started 2 tasks: (task 6b2b6a2d-a0ca-4904-a971-76895c65f3ae) (task f416f133-6e8e-4872-852f-64d0668619ae).
Deployment unready
Deployment Desired Pending Running
PRIMARY 2 2 0
Service:
(service app) has started 2 tasks: (task 6b2b6a2d-a0ca-4904-a971-76895c65f3ae) (task f416f133-6e8e-4872-852f-64d0668619ae).
Deployment unready
When I enable service discovery my tasks go into "activating" status and then stop, get recreated again, etc.
Not entirely sure how to fix, but I was doing some digging and this might be of help. Adding some extra config around healthchecks:
https://forums.aws.amazon.com/thread.jspa?threadID=283572