-
Notifications
You must be signed in to change notification settings - Fork 24
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
Add ECS Workload Agent #725
Add ECS Workload Agent #725
Conversation
Note to self: always, always, always rebuild after rebasing. |
cea6e9a
to
43d48ca
Compare
Still testing to be done, but marking this as ready for review since I think the error is something in my local test setup and not necessarily anything with the agent itself. |
0e19fef
to
1ca83cb
Compare
1ca83cb
to
8dadb10
Compare
8dadb10
to
9e1fb37
Compare
9e1fb37
to
6f732ed
Compare
This adds a new test agent for running workload testing in an ECS cluster. Signed-off-by: Sean McGinnis <[email protected]>
6f732ed
to
b892b1e
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM. Can you verify that the nvidia test works with this agent?
Able to deploy an ECS cluster instance running Bottlerocket 1.11.1, apiVersion: testsys.system/v1
kind: Test
metadata:
name: ecs-workload
namespace: testsys
spec:
agent:
configuration:
region: "us-east-2"
clusterName: "brtest"
tests:
- name: nvidia-workload
image: public.ecr.aws/p0f9g6e8/gpu-tests:latest
gpu: true
image: ecs-workload-agent:demo
name: ecs-workload-test-agent
keepRunning: true
timeout: "5000"
secrets:
awsCredentials: "aws-creds"
resources:
- brtest-instances
- brtest
dependsOn: []
---
apiVersion: testsys.system/v1
kind: Resource
metadata:
name: brtest
namespace: testsys
spec:
agent:
name: ecs-provider
image: ecs-resource-agent:demo
keepRunning: true
configuration:
clusterName: brtest
region: "us-east-2"
secrets:
awsCredentials: "aws-creds"
dependsOn: []
destructionPolicy: never
---
apiVersion: testsys.system/v1
kind: Resource
metadata:
name: brtest-instances
namespace: testsys
spec:
agent:
name: ec2-provider
image: ec2-resource-agent:demo
keepRunning: true
configuration:
clusterName: brtest
clusterType: ecs
instanceCount: 2
instanceProfileArn: ${brtest.iamInstanceProfileArn}
nodeAmi: "ami-0ba78cb33c07e2333"
region: "us-east-2"
subnetIds: ${brtest.publicSubnetIds}
instanceTypes: ["g4dn.xlarge"]
secrets:
awsCredentials: "aws-creds"
dependsOn: [brtest]
destructionPolicy: never Able to create cluster, launch and join EC2 instances, and execute NVIDIA workload test. Logs from the workload test: $ k -n testsys logs pod/ecs-workload-qzxnr
[2023-01-09T16:27:02Z INFO ecs_workload_agent] Initializing ECS workload test agent...
[2023-01-09T16:27:02Z INFO ecs_workload_agent] Waiting for registered container instances...
[2023-01-09T16:27:02Z INFO ecs_workload_agent] Waiting for cluster to have registered instances
[2023-01-09T16:27:04Z INFO ecs_workload_agent] Waiting for cluster to have registered instances
[2023-01-09T16:27:06Z INFO ecs_workload_agent] Waiting for cluster to have registered instances
[2023-01-09T16:27:08Z INFO ecs_workload_agent] Waiting for cluster to have registered instances
[2023-01-09T16:27:09Z INFO ecs_workload_agent] Running task 'arn:aws:ecs:us-east-2:861807767978:task-definition/testsys-bottlerocket-nvidia-workload:1'
[2023-01-09T16:27:09Z INFO ecs_workload_agent] Waiting for tasks to complete...
[2023-01-09T16:27:57Z INFO test_agent::agent] Test execution finished without returning an error.
[2023-01-09T16:27:57Z INFO test_agent::agent] Test output tarball created.
[2023-01-09T16:27:57Z INFO test_agent::agent] 'keep_running' is true. Stopped task in the ECS cluster shows exited with exit code 0. |
Issue number:
Closes: #617
Description of changes:
This adds a new test agent for running workload testing in an ECS cluster.
Testing done:
See details of testing in comment below
Terms of contribution:
By submitting this pull request, I agree that this contribution is dual-licensed under the terms of both the Apache License, version 2.0, and the MIT license.