Implement build-and-test Github Actions CI using AWS spot instances

.github/workflows/build-and-test-x86_64.yml

@@ -0,0 +1,11 @@
+name: Build And Test (x86-64)
+
+on: [push, pull_request]
+
+jobs:
+  build-and-test-x86-64:
+    uses: ./.github/workflows/build-and-test.yml
+    with:
+      architecture: x86_64
+      instance_type: c5.9xlarge
+    secrets: inherit

.github/workflows/build-and-test.yml

@@ -0,0 +1,150 @@
+name: Build And Test
+
+# Based on
+# https://blog.devgenius.io/create-register-and-terminate-ec2-spot-instances-via-github-actions-aef473b8a00f
+
+on:
+  workflow_call:
+    inputs:
+      architecture:
+        required: true
+        type: string
+      instance_type:
+        required: true
+        type: string
+    secrets:
+      ACTIONS_TOKEN:
+        required: true
+
+env:
+  EC2_SECURITY_GROUP_ID: sg-072c6deccd49e89c8 # rr-testing
+  AWS_ROLE_ARN: arn:aws:iam::019859450731:role/SpotInstanceManager
+  AWS_REGION: us-east-2
+  RUNNER_VERSION: "2.316.1"
+  RUNNER_LABEL: ${{ inputs.architecture }}_${{github.sha}}
+
+jobs:
+  start-runner:
+    name: Start Runner
+    permissions:
+      id-token: write
+    runs-on: ubuntu-latest
+    outputs:
+      EC2_HOST: ${{ env.EC2_HOST }}
+      EC2_INSTANCE_ID: ${{ env.EC2_INSTANCE_ID }}
+    steps:
+      - name: "Run :: Configure AWS credentials"
+        uses: aws-actions/configure-aws-credentials@v4
+        with:
+          role-to-assume: ${{ env.AWS_ROLE_ARN }}
+          role-session-name: pipeline-assume-role-session
+          aws-region: ${{ env.AWS_REGION }}
+
+      - name: "Run :: Create and register AWS spot instance"
+        run: |2-
+              # Get GitHub Actions runner registration token
+              response=$(curl -s -X POST -H 'Accept: application/vnd.github+json' -H 'Authorization: Bearer ${{ secrets.ACTIONS_TOKEN }}' -H 'X-GitHub-Api-Version: 2022-11-28' https://api.github.com/orgs/rr-debugger/actions/runners/registration-token)
+              RUNNER_REG_TOKEN=$(echo "$response" | jq -r .token)
+              echo "RUNNER_REG_TOKEN :: $RUNNER_REG_TOKEN"
+
+              if [ $RUNNER_REG_TOKEN == "null" ]; then
+                exit 1
+              fi
+
+              USER_DATA="#!/bin/bash
+
+              # Make sure the VM doesn't run for more than an hour
+              shutdown +60
+
+              apt-get update -y
+              apt-get dist-upgrade -f -y
+                
+              sudo -u ubuntu --login <<EOF
+              # Install GitHub Actions runner
+              mkdir /home/ubuntu/actions-runner && cd /home/ubuntu/actions-runner
+              curl -o actions-runner-linux-x64-${RUNNER_VERSION}.tar.gz -L https://github.com/actions/runner/releases/download/v${RUNNER_VERSION}/actions-runner-linux-x64-${RUNNER_VERSION}.tar.gz
+              echo \"Github Runner Installed\"
+              # Extract the installer
+              tar xzf ./actions-runner-linux-x64-${RUNNER_VERSION}.tar.gz
+              echo \"Github Runner Installer Extracted\"
+                
+              # Run GitHub Actions runner configuration
+              yes '' | ./config.sh --url https://github.com/rr-debugger --token $RUNNER_REG_TOKEN --labels ${{ env.RUNNER_LABEL }}
+
+              # Run GitHub Actions runner
+              yes '' | ./run.sh 
+              EOF
+              "
+                
+              IMAGE_ID=$(aws ec2 describe-images --owners 099720109477 --filters 'Name=architecture,Values=${{ inputs.architecture }}' 'Name=name,Values=ubuntu/images/hvm-ssd-gp3/ubuntu-noble-24.04-*' | jq --raw-output .Images[0].ImageId)
+              echo "IMAGE_ID :: $IMAGE_ID"
+
+              ENCODED_USER_DATA=$(echo -n "$USER_DATA" | sed 's/^[[:space:]]*//' | base64 -w 0)
+                
+              INSTANCE_ID=$(aws ec2 run-instances --image-id $IMAGE_ID --instance-type ${{ inputs.instance_type }} --user-data $ENCODED_USER_DATA --security-group-ids ${{ env.EC2_SECURITY_GROUP_ID }} --instance-market-options MarketType=spot --instance-initiated-shutdown-behavior terminate --query 'Instances[0].InstanceId' --key-name rr-testing --output text)
+              echo "INSTANCE_ID :: $INSTANCE_ID"
+              aws ec2 wait instance-running --instance-ids $INSTANCE_ID
+
+              HOSTNAME=$(aws ec2 describe-instances --instance-ids $INSTANCE_ID --query 'Reservations[0].Instances[0].PublicDnsName' --output text)
+              echo "HOSTNAME :: $HOSTNAME"
+
+              echo RUNNER_LABEL :: ${{ env.RUNNER_LABEL }}
+
+              echo "EC2_HOST=$HOSTNAME" >> $GITHUB_ENV
+              echo "EC2_INSTANCE_ID=$INSTANCE_ID" >> $GITHUB_ENV 
+
+  build-and-test:
+    name: Build And Test
+    permissions:
+      contents: read
+    runs-on:
+      - self-hosted
+      - ${{ inputs.architecture }}_${{github.sha}}
+    needs:
+      - start-runner
+    steps:
+      - name: Start
+        run: |2-
+              echo " Starting GitHub Action!" && 
+              echo "STEPS_CAN_PROCEED=true" >> $GITHUB_ENV
+
+      - name: "Run :: Checkout repository"
+        uses: actions/checkout@v2
+
+      - name: "Run :: Build"
+        run: ./scripts/github-actions-build.sh
+
+      - name: "Run :: Test"
+        run: ./scripts/github-actions-test.sh
+
+  stop-runner:
+    name: Stop Runner
+    permissions:
+      id-token: write
+    runs-on: ubuntu-latest
+    needs:
+      - start-runner
+      - build-and-test
+    if: ${{ always() }}
+    steps:
+      - name: "Run :: Configure AWS credentials"
+        uses: aws-actions/configure-aws-credentials@v4
+        with:
+          role-to-assume: ${{ env.AWS_ROLE_ARN }}
+          role-session-name: pipeline-assume-role-session
+          aws-region: ${{ env.AWS_REGION }}
+
+      - name: "Run :: Terminate and unregister AWS spot instance"
+        run: |2-
+              aws ec2 terminate-instances --instance-ids ${{ needs.start-runner.outputs.EC2_INSTANCE_ID }}
+              gh_runner_label=${{ needs.start-runner.outputs.EC2_RUNNER }}
+              response=$(curl -X GET https://api.github.com/orgs/rr-debugger/actions/runners -H 'Authorization: Bearer ${{ secrets.ACTIONS_TOKEN }}'
+              )
+              offline_runners=$(echo "$response" | jq '.runners | map(select((.labels? | any(.name == "${{ inputs.architecture }}_${{github.sha}}"))))')
+              echo "Offline Runners :: $offline_runners"
+              
+              echo "Attempting to remove offline runners..."
+              for runner in $(echo "$offline_runners" | jq -r '.[].id'); do
+                echo "Triggering action for runner ID: $runner"
+                curl -X DELETE https://api.github.com/orgs/rr-debugger/actions/runners/$runner -H 'Authorization: Bearer ${{ secrets.ACTIONS_TOKEN }}'
+              done

scripts/github-actions-build.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+set +x # echo commands
+set -e # default to exiting on error"
+
+uname -a
+
+sudo apt-get install -y rpm ccache cmake g++ pkg-config zlib1g-dev git python-dev-is-python3 libacl1-dev ninja-build manpages-dev capnproto libcapnp-dev gdb lldb python3-pexpect
+
+mkdir obj
+cd obj
+cmake -G Ninja -DCMAKE_BUILD_TYPE=DEBUG -Dstaticlibs=FALSE ..
+ninja

scripts/github-actions-test.sh

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+set +x # echo commands
+set -e # default to exiting on error"
+
+# Enable perf events for rr
+echo 0 | sudo tee /proc/sys/kernel/perf_event_paranoid
+# Enable ptrace-attach to any process. This lets us get more data when tests fail.
+echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope
+# Disable AppArmor restrictions on user namespaces, which our tests need to use
+(echo 0 | sudo tee /proc/sys/kernel/apparmor_restrict_unprivileged_userns) || true
+let halfproc=`nproc`/2
+ctest -j$halfproc --verbose