Skip to content

Fix issue with AWS GPU runners not picking up jobs. #107

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.

Sign up for GitHub

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

Closed
WilliamYue37 wants to merge 1 commit into from
Closed

Fix issue with AWS GPU runners not picking up jobs. #107

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 21 additions & 2 deletions .github/workflows/gpu_test.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,6 +30,7 @@ jobs:
start-runner:
name: Start GPU Runner
runs-on: ubuntu-latest
timeout-minutes: 15 # Fail fast if ASG/runner never becomes ready
permissions:
id-token: write # Required for requesting the JWT
contents: read
Expand All @@ -42,8 +43,26 @@ jobs:

- name: Start Instance
run: |
aws autoscaling set-desired-capacity --auto-scaling-group-name github-runner-asg-g6-2xlarge --desired-capacity 1
echo "Waiting for instance to be ready..."
ASG_NAME="github-runner-asg-g6-2xlarge"
aws autoscaling set-desired-capacity --auto-scaling-group-name "$ASG_NAME" --desired-capacity 1
echo "Waiting for instance to be InService (max 10 min)..."
for i in $(seq 1 40); do
COUNT=$(aws autoscaling describe-auto-scaling-groups --auto-scaling-group-names "$ASG_NAME" \
--query 'AutoScalingGroups[0].Instances[?LifecycleState==`InService`].InstanceId' --output text | wc -w)
[ "$COUNT" -ge 1 ] && break
echo "Instance not ready yet ($i/40)..."
sleep 15
done
if [ "$COUNT" -lt 1 ]; then
echo "Timeout: instance did not become InService"
exit 1
fi
INSTANCE_ID=$(aws autoscaling describe-auto-scaling-groups --auto-scaling-group-names "$ASG_NAME" \
--query 'AutoScalingGroups[0].Instances[?LifecycleState==`InService`].InstanceId' --output text | awk '{print $1}')
echo "Waiting for EC2 status checks on $INSTANCE_ID..."
aws ec2 wait instance-status-ok --instance-ids "$INSTANCE_ID"
echo "Instance ready. Waiting 3 min for GitHub runner to register..."
sleep 180

gpu-test:
name: Run Pytest on GPU
Expand Down
23 changes: 21 additions & 2 deletions .github/workflows/regression_test.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,7 @@ jobs:
start-runner:
name: Start GPU Runner
runs-on: ubuntu-latest
timeout-minutes: 15 # Fail fast if ASG/runner never becomes ready
permissions:
id-token: write
contents: read
Expand All @@ -44,8 +45,26 @@ jobs:

- name: Start Instance
run: |
aws autoscaling set-desired-capacity --auto-scaling-group-name github-runner-asg-g6-12xlarge --desired-capacity 1
echo "Waiting for instance to be ready..."
ASG_NAME="github-runner-asg-g6-12xlarge"
aws autoscaling set-desired-capacity --auto-scaling-group-name "$ASG_NAME" --desired-capacity 1
echo "Waiting for instance to be InService (max 10 min)..."
for i in $(seq 1 40); do
COUNT=$(aws autoscaling describe-auto-scaling-groups --auto-scaling-group-names "$ASG_NAME" \
--query 'AutoScalingGroups[0].Instances[?LifecycleState==`InService`].InstanceId' --output text | wc -w)
[ "$COUNT" -ge 1 ] && break
echo "Instance not ready yet ($i/40)..."
sleep 15
done
if [ "$COUNT" -lt 1 ]; then
echo "Timeout: instance did not become InService"
exit 1
fi
INSTANCE_ID=$(aws autoscaling describe-auto-scaling-groups --auto-scaling-group-names "$ASG_NAME" \
--query 'AutoScalingGroups[0].Instances[?LifecycleState==`InService`].InstanceId' --output text | awk '{print $1}')
echo "Waiting for EC2 status checks on $INSTANCE_ID..."
aws ec2 wait instance-status-ok --instance-ids "$INSTANCE_ID"
echo "Instance ready. Waiting 3 min for GitHub runner to register..."
sleep 180

train-regression:
name: Train with Model Parallelism
Expand Down
Loading