zebra/.github/workflows/test.yml

388 lines
20 KiB
YAML

name: Test
on:
workflow_dispatch:
inputs:
network:
default: 'Mainnet'
regenerate-disks:
type: boolean
default: false
description: Just update stateful disks
pull_request:
branches:
- main
paths:
- '**/*.rs'
- '**/*.txt'
- '**/Cargo.toml'
- '**/Cargo.lock'
- 'docker/**'
- '.github/workflows/test.yml'
env:
CARGO_INCREMENTAL: '1'
ZEBRA_SKIP_IPV6_TESTS: "1"
NETWORK: Mainnet
PROJECT_ID: ${{ secrets.GCP_PROJECT_ID }}
GAR_BASE: us-docker.pkg.dev/${{ secrets.GCP_PROJECT_ID }}/zebra
GCR_BASE: gcr.io/${{ secrets.GCP_PROJECT_ID }}
REGION: us-central1
ZONE: us-central1-a
MACHINE_TYPE: c2-standard-8
IMAGE_NAME: zebrad-test
jobs:
build:
name: Build images
timeout-minutes: 210
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v2.4.0
with:
persist-credentials: false
- name: Inject slug/short variables
uses: rlespinasse/github-slug-action@v4
# Setup Docker Buildx to allow use of docker cache layers from GH
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v1
with:
driver-opts: network=host
- name: Login to Google Artifact Registry
uses: docker/login-action@v1.12.0
with:
registry: us-docker.pkg.dev
username: _json_key
password: ${{ secrets.GOOGLE_CREDENTIALS }}
- name: Login to Google Container Registry
uses: docker/login-action@v1.12.0
with:
registry: gcr.io
username: _json_key
password: ${{ secrets.GOOGLE_CREDENTIALS }}
# Build and push image to Google Artifact Registry
- name: Build & push
id: docker_build
uses: docker/build-push-action@v2.8.0
with:
target: tester
context: .
file: ./docker/Dockerfile.build
tags: |
${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:latest
${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}
${{ env.GCR_BASE }}/${{ env.GITHUB_REPOSITORY_SLUG_URL }}/${{ env.IMAGE_NAME }}:latest
${{ env.GCR_BASE }}/${{ env.GITHUB_REPOSITORY_SLUG_URL }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}
build-args: |
NETWORK=${{ github.event.inputs.network || env.NETWORK }}
SHORT_SHA=${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}
RUST_BACKTRACE=full
ZEBRA_SKIP_NETWORK_TESTS="1"
CHECKPOINT_SYNC=${{ github.event.inputs.checkpoint_sync || true }}
RUST_LOG=debug
SENTRY_DSN=${{ secrets.SENTRY_ENDPOINT }}
push: true
cache-from: type=gha
cache-to: type=gha,mode=max
# Run all the zebra tests, including tests that are ignored by default
test-all:
name: Test all
runs-on: ubuntu-latest
needs: build
if: ${{ github.event.inputs.regenerate-disks != 'true' }}
steps:
- uses: actions/checkout@v2.4.0
with:
persist-credentials: false
- name: Inject slug/short variables
uses: rlespinasse/github-slug-action@v4
- name: Run all zebrad tests
run: |
docker pull ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}
docker run -e ZEBRA_SKIP_IPV6_TESTS --name zebrad-tests -t ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} cargo test --locked --release --features enable-sentry --workspace -- --include-ignored
test-fake-activation-heights:
name: Test with fake activation heights
runs-on: ubuntu-latest
needs: build
if: ${{ github.event.inputs.regenerate-disks != 'true' }}
steps:
- uses: actions/checkout@v2.4.0
with:
persist-credentials: false
- name: Inject slug/short variables
uses: rlespinasse/github-slug-action@v4
- name: Run tests with fake activation heights
run: |
docker pull ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}
docker run -e ZEBRA_SKIP_IPV6_TESTS --name zebrad-tests -t ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} cargo test --locked --release --package zebra-state --lib -- with_fake_activation_heights
# Test that Zebra syncs and checkpoints a few thousand blocks from an empty state
test-empty-sync:
name: Test checkpoint sync from empty state
runs-on: ubuntu-latest
needs: build
if: ${{ github.event.inputs.regenerate-disks != 'true' }}
steps:
- uses: actions/checkout@v2.4.0
with:
persist-credentials: false
- name: Inject slug/short variables
uses: rlespinasse/github-slug-action@v4
- name: Run zebrad large sync tests
run: |
docker pull ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}
docker run -e ZEBRA_SKIP_IPV6_TESTS --name zebrad-tests -t ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} cargo test --locked --release --features enable-sentry --test acceptance sync_large_checkpoints_ -- --ignored
regenerate-stateful-disks:
name: Regenerate stateful disks
runs-on: ubuntu-latest
needs: build
outputs:
disk_short_sha: ${{ steps.disk-short-sha.outputs.disk_short_sha }}
steps:
- uses: actions/checkout@v2.4.0
with:
persist-credentials: false
fetch-depth: '2'
- name: Get specific changed files
id: changed-files-specific
uses: tj-actions/changed-files@v14.4
with:
files: |
/zebra-state/**/disk_format.rs
/zebra-state/**/disk_db.rs
/zebra-state/**/finalized_state.rs
/zebra-state/**/constants.rs
- name: Inject slug/short variables
uses: rlespinasse/github-slug-action@v4
- name: Downcase network name for disks
run: |
echo LOWER_NET_NAME="${{ github.event.inputs.network || env.NETWORK }}" | awk '{print tolower($0)}' >> $GITHUB_ENV
# Setup gcloud CLI
- name: Authenticate to Google Cloud
id: auth
uses: google-github-actions/auth@v0.5.0
with:
credentials_json: ${{ secrets.GOOGLE_CREDENTIALS }}
- name: Create GCP compute instance
id: create-instance
if: ${{ steps.changed-files-specific.outputs.any_changed == 'true' || github.event.inputs.regenerate-disks == 'true' }}
run: |
gcloud compute instances create-with-container "zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}" \
--boot-disk-size 100GB \
--boot-disk-type pd-ssd \
--create-disk name="zebrad-cache-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-${{ env.lower_net_name }}-canopy",size=100GB,type=pd-ssd \
--container-mount-disk mount-path='/zebrad-cache',name="zebrad-cache-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-${{ env.lower_net_name }}-canopy" \
--container-image ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} \
--container-restart-policy=never \
--container-stdin \
--container-tty \
--container-command="cargo" \
--container-arg="test" \
--container-arg="--locked" \
--container-arg="--release" \
--container-arg="--features" \
--container-arg="enable-sentry,test_sync_to_mandatory_checkpoint_${{ env.lower_net_name }}" \
--container-arg="--manifest-path" \
--container-arg="zebrad/Cargo.toml" \
--container-arg="sync_to_mandatory_checkpoint_${{ env.lower_net_name }}" \
--container-env=ZEBRA_SKIP_IPV6_TESTS=1 \
--machine-type ${{ env.MACHINE_TYPE }} \
--scopes cloud-platform \
--metadata=google-monitoring-enabled=true,google-logging-enabled=true \
--tags zebrad \
--zone "${{ env.ZONE }}"
# TODO: this approach is very mesy, but getting the just created container name is very error prone and GCP doesn't have a workaround for this without requiring a TTY
# This TODO relates to the following issues:
# https://github.com/actions/runner/issues/241
# https://www.googlecloudcommunity.com/gc/Infrastructure-Compute-Storage/SSH-into-Compute-Container-not-easily-possible/td-p/170915
- name: Get container name from logs
id: get-container-name
if: steps.create-instance.outcome == 'success'
run: |
INSTANCE_ID=$(gcloud compute instances describe zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} --zone ${{ env.ZONE }} --format='value(id)')
echo "Using instance: $INSTANCE_ID"
while [[ ${CONTAINER_NAME} != *"zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}"* ]]; do
CONTAINER_NAME=$(gcloud logging read 'log_name=projects/${{ env.PROJECT_ID }}/logs/cos_system AND jsonPayload.MESSAGE:zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}' --format='value(jsonPayload.MESSAGE)' --limit=1 | grep -o '...-zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-....' | tr -d "'.")
echo "Using container: ${CONTAINER_NAME} from instance: ${INSTANCE_ID}"
sleep 10
done
CONTAINER_NAME=$(gcloud logging read 'log_name=projects/${{ env.PROJECT_ID }}/logs/cos_system AND jsonPayload.MESSAGE:zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}' --format='value(jsonPayload.MESSAGE)' --limit=1 | grep -o '...-zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-....' | tr -d "'.")
echo "::set-output name=zebra_container::$CONTAINER_NAME"
- name: Regenerate stateful disks logs
id: sync-to-checkpoint
if: steps.create-instance.outcome == 'success'
run: |
gcloud compute ssh \
zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} \
--zone ${{ env.ZONE }} \
--quiet \
--ssh-flag="-o ServerAliveInterval=5" \
--command="docker logs --follow ${{ env.ZEBRA_CONTAINER }}"
env:
ZEBRA_CONTAINER: ${{ steps.get-container-name.outputs.zebra_container }}
# Create image from disk that will be used to sync past mandatory checkpoint test
# Force the image creation as the disk is still attached even though is not being used by the container
- name: Create image from state disk
# Only run if the earlier step succeeds
if: steps.sync-to-checkpoint.outcome == 'success'
run: |
gcloud compute images create zebrad-cache-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-${{ env.lower_net_name }}-canopy \
--force \
--source-disk=zebrad-cache-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-${{ env.lower_net_name }}-canopy \
--source-disk-zone=${{ env.ZONE }} \
--storage-location=us \
--description="Created from head branch ${{ env.GITHUB_HEAD_REF_SLUG_URL }} targeting ${{ env.GITHUB_BASE_REF_SLUG }} from PR ${{ env.GITHUB_REF_SLUG_URL }} with commit ${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA }}"
- name: Output and write the disk SHORT_SHA to a txt
id: disk-short-sha
if: steps.sync-to-checkpoint.outcome == 'success'
run: |
short_sha=$(echo "${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}")
echo "$short_sha" > latest-disk-state-sha.txt
echo "::set-output name=disk_short_sha::$short_sha"
- name: Upload the disk state txt
if: steps.sync-to-checkpoint.outcome == 'success'
uses: actions/upload-artifact@v2.3.1
with:
name: latest-disk-state-sha
path: latest-disk-state-sha.txt
retention-days: 1095
- name: Delete test instance
# Do not delete the instance if the sync timeouts in GitHub
if: ${{ steps.sync-to-checkpoint.outcome == 'success' }} || ${{ steps.sync-to-checkpoint.outcome == 'failure' }}
continue-on-error: true
run: |
gcloud compute instances delete "zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}" --delete-disks all --zone "${{ env.ZONE }}"
# Test that Zebra syncs and fully validates a few thousand blocks from a cached post-checkpoint state
test-stateful-sync:
name: Test full validation sync from cached state
runs-on: ubuntu-latest
needs: [ build, regenerate-stateful-disks]
steps:
- uses: actions/checkout@v2.4.0
with:
persist-credentials: false
- name: Inject slug/short variables
uses: rlespinasse/github-slug-action@v4
- name: Downcase network name for disks
run: |
echo LOWER_NET_NAME="${{ github.event.inputs.network || env.NETWORK }}" | awk '{print tolower($0)}' >> $GITHUB_ENV
# Get the latest uploaded txt with the disk SHORT_SHA from this workflow
- name: Download latest disk state SHORT_SHA
uses: dawidd6/action-download-artifact@v2.17.0
with:
github_token: ${{ secrets.GITHUB_TOKEN }}
workflow: test.yml
workflow_conclusion: ''
name: latest-disk-state-sha
check_artifacts: true
- name: Get disk state SHA from txt
id: get-disk-sha
run: |
output=$(cat latest-disk-state-sha.txt)
echo "::set-output name=sha::$output"
# Setup gcloud CLI
- name: Authenticate to Google Cloud
id: auth
uses: google-github-actions/auth@v0.5.0
with:
credentials_json: ${{ secrets.GOOGLE_CREDENTIALS }}
# Creates Compute Engine virtual machine instance w/ disks
- name: Create GCP compute instance
id: create-instance
run: |
gcloud compute instances create-with-container "zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}" \
--boot-disk-size 100GB \
--boot-disk-type pd-ssd \
--create-disk=image=zebrad-cache-${{ env.DISK_SHORT_SHA }}-${{ env.lower_net_name }}-canopy,name=zebrad-cache-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-${{ env.lower_net_name }}-canopy,size=100GB,type=pd-ssd \
--container-mount-disk=mount-path='/zebrad-cache',name=zebrad-cache-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-${{ env.lower_net_name }}-canopy \
--container-image ${{ env.GAR_BASE }}/${{ env.IMAGE_NAME }}:${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} \
--container-restart-policy=never \
--container-stdin \
--container-tty \
--container-command="cargo" \
--container-arg="test" \
--container-arg="--locked" \
--container-arg="--release" \
--container-arg="--features" \
--container-arg="enable-sentry,test_sync_past_mandatory_checkpoint_${{ env.lower_net_name }}" \
--container-arg="--manifest-path" \
--container-arg="zebrad/Cargo.toml" \
--container-arg="sync_past_mandatory_checkpoint_${{ env.lower_net_name }}" \
--container-env=ZEBRA_SKIP_IPV6_TESTS=1 \
--machine-type ${{ env.MACHINE_TYPE }} \
--scopes cloud-platform \
--metadata=google-monitoring-enabled=true,google-logging-enabled=true \
--tags zebrad \
--zone "${{ env.ZONE }}"
env:
DISK_SHORT_SHA: ${{ needs.regenerate-stateful-disks.outputs.disk_short_sha || steps.get-disk-sha.outputs.sha }}
# TODO: this approach is very mesy, but getting the just created container name is very error prone and GCP doesn't have a workaround for this without requiring a TTY
# This TODO relates to the following issues:
# https://github.com/actions/runner/issues/241
# https://www.googlecloudcommunity.com/gc/Infrastructure-Compute-Storage/SSH-into-Compute-Container-not-easily-possible/td-p/170915
- name: Get container name from logs
id: get-container-name
if: steps.create-instance.outcome == 'success'
run: |
INSTANCE_ID=$(gcloud compute instances describe zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} --zone ${{ env.ZONE }} --format='value(id)')
echo "Using instance: $INSTANCE_ID"
while [[ ${CONTAINER_NAME} != *"zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}"* ]]; do
CONTAINER_NAME=$(gcloud logging read 'log_name=projects/${{ env.PROJECT_ID }}/logs/cos_system AND jsonPayload.MESSAGE:zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}' --format='value(jsonPayload.MESSAGE)' --limit=1 | grep -o '...-zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-....' | tr -d "'.")
echo "Using container: ${CONTAINER_NAME} from instance: ${INSTANCE_ID}"
sleep 10
done
CONTAINER_NAME=$(gcloud logging read 'log_name=projects/${{ env.PROJECT_ID }}/logs/cos_system AND jsonPayload.MESSAGE:zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}' --format='value(jsonPayload.MESSAGE)' --limit=1 | grep -o '...-zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}-....' | tr -d "'.")
echo "::set-output name=zebra_container::$CONTAINER_NAME"
- name: Sync past mandatory checkpoint logs
id: sync-past-checkpoint
run: |
gcloud compute ssh \
zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }} \
--zone ${{ env.ZONE }} \
--quiet \
--ssh-flag="-o ServerAliveInterval=5" \
--command="docker logs --follow ${{ env.ZEBRA_CONTAINER }}"
env:
ZEBRA_CONTAINER: ${{ steps.get-container-name.outputs.zebra_container }}
- name: Delete test instance
# Do not delete the instance if the sync timeouts in GitHub
if: ${{ steps.sync-past-checkpoint.outcome == 'success' }} || ${{ steps.sync-past-checkpoint.outcome == 'failure' }}
continue-on-error: true
run: |
gcloud compute instances delete "zebrad-tests-${{ env.GITHUB_HEAD_REF_SLUG_URL || env.GITHUB_REF_SLUG_URL }}-${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT || env.GITHUB_SHA_SHORT }}" --delete-disks all --zone "${{ env.ZONE }}"