daml/azure-cron.yml
Gary Verhaegen 4a6ab84b69
add default machine capability (#5912)
add default machine capability

We semi-regularly need to do work that has the potential to disrupt a
machine's local cache, rendering it broken for other streams of work.
This can include upgrading nix, upgrading Bazel, debugging caching
issues, or anything related to Windows.

Right now we do not have any good solution for these situations. We can
either not do those streams of work, or we can proceed with them and
just accept that all other builds may get affected depending on which
machine they get assigned to. Debugging broken nodes is particularly
tricky as we do not have any way to force a build to run on a given
node.

This PR aims at providing a better alternative by (ab)using an Azure
Pipelines feature called
[capabilities](https://docs.microsoft.com/en-us/azure/devops/pipelines/agents/agents?view=azure-devops&tabs=browser#capabilities).
The idea behind capabilities is that you assign a set of tags to a
machine, and then a job can express its
[demands](https://docs.microsoft.com/en-us/azure/devops/pipelines/process/demands?view=azure-devops&tabs=yaml),
i.e. specify a set of tags machines need to have in order to run it.

Support for this is fairly badly documented. We can gather from the
documentation that a job can specify two things about a capability
(through its `demands`): that a given tag exists, and that a given tag
has an exact specified value. In particular, a job cannot specify that a
capability should _not_ be present, meaning we cannot rely on, say,
adding a "broken" tag to broken machines.

Documentation on how to set capabilities for an agent is basically
nonexistent, but [looking at the
code](https://github.com/microsoft/azure-pipelines-agent/blob/master/src/Microsoft.VisualStudio.Services.Agent/Capabilities/UserCapabilitiesProvider.cs)
indicates that they can be set by using a simple `key=value`-formatted
text file, provided we can find the right place to put this file.

This PR adds this file to our Linux, macOS and Windows node init scripts
to define an `assignment` capability and adds a demand for a `default`
value on each job. From then on, when we hit a case where we want a PR
to run on a specific node, and to prevent other PRs from running on that
node, we can manually override the capability from the Azure UI and
update the demand in the relevant YAML file in the PR.

CHANGELOG_BEGIN
CHANGELOG_END
2020-05-09 18:21:42 +02:00

189 lines
7.7 KiB
YAML

# Copyright (c) 2020 Digital Asset (Switzerland) GmbH and/or its affiliates. All rights reserved.
# SPDX-License-Identifier: Apache-2.0
# Azure Pipelines file, see https://aka.ms/yaml
# Do not run on PRs
pr: none
# Do not run on merge to master
trigger: none
# Do run on a schedule (hourly)
#
# This is currently (2019-08-15) broken on Azure for GitHub-hosted repos. It
# does, however, work as expected for Azure-hosted repos. As a workaround, we
# have created a repo inside Azure that contains an `azure-pipelines.yml` file
# that just triggers this job.
#
# When the situation is resolved, delete that repo in Azure and uncomment the
# following. In the meantime, this should stay commented so we avoid running
# jobs twice when Azure fixes this issue.
#schedules:
#- cron: "0 * * * *"
# displayName: hourly cron
# branches:
# include:
# - master
# always: true
jobs:
- job: docs
timeoutInMinutes: 120
pool:
name: 'linux-pool'
demands: assignment -equals default
steps:
- checkout: self
- bash: ci/dev-env-install.sh
displayName: 'Build/Install the Developer Environment'
- bash: ci/cron/run.sh
env:
AWS_ACCESS_KEY_ID: $(AWS_ACCESS_KEY_ID)
AWS_SECRET_ACCESS_KEY: $(AWS_SECRET_ACCESS_KEY)
- bash: |
set -euo pipefail
MESSAGE=$(git log --pretty=format:%s -n1)
curl -XPOST \
-i \
-H 'Content-type: application/json' \
--data "{\"text\":\"<!here> *FAILED* Daily Docs: <https://dev.azure.com/digitalasset/daml/_build/results?buildId=$(Build.BuildId)|$MESSAGE>\n\"}" \
$(Slack.team-daml)
condition: and(failed(), eq(variables['Build.SourceBranchName'], 'master'))
- template: ci/tell-slack-failed.yml
- job: docker_image
timeoutInMinutes: 60
pool:
name: 'linux-pool'
demands: assignment -equals default
steps:
- checkout: self
- bash: |
set -euo pipefail
eval "$(dev-env/bin/dade-assist)"
HEAD=$(git rev-parse HEAD)
while ! nix-build -A tools.sed -A tools.jq -A tools.curl nix; do :; done
echo $DOCKER_PASSWORD | docker login --username $DOCKER_LOGIN --password-stdin
RELEASES=$(curl https://api.github.com/repos/digital-asset/daml/releases -s | jq -r '.[] | .tag_name')
DIR=$(pwd)
VERSIONS=$(curl 'https://hub.docker.com/v2/repositories/digitalasset/daml-sdk/tags/?page_size=10000' -s)
# Our docker tags should be stable. Therefore, we only build the image if it has not already
# been built before and we checkout the Dockerfile for the release tag.
# We do not update docker images for older releases so only docker images for SDK releases
# >= 0.13.43 are built this way.
for version in $(echo $RELEASES | sed -e 's/ /\n/g'); do
LAST_UPDATE=$(echo $VERSIONS | jq -r '.results[] | select(.name == "'${version#v}'") | .last_updated')
if [[ -n "$LAST_UPDATE" ]]; then
echo "${version#v} already exists, skipping."
else
echo "Building version ${version#v}..."
git checkout "$version"
cd ci/docker/daml-sdk
docker build -t digitalasset/daml-sdk:${version#v} --build-arg VERSION=${version#v} .
docker push digitalasset/daml-sdk:${version#v}
cd "$DIR"
git checkout $HEAD
echo "Done."
fi
done
env:
DOCKER_LOGIN: $(DOCKER_LOGIN)
DOCKER_PASSWORD: $(DOCKER_PASSWORD)
- template: ci/tell-slack-failed.yml
- job: vscode_marketplace
timeoutInMinutes: 10
pool:
name: 'linux-pool'
demands: assignment -equals default
steps:
- checkout: self
- bash: |
set -euo pipefail
eval "$(dev-env/bin/dade-assist)"
AUTH=$(echo -n "OAuth:${MARKETPLACE_TOKEN}" | base64 -w0)
MARKET=$(curl -H "Authorization: Basic $AUTH" \
-H "Accept: application/json;api-version=5.0-preview.2" \
-s \
"https://marketplace.visualstudio.com/_apis/gallery/publishers/DigitalAssetHoldingsLLC/extensions/daml?flags=1" \
| jq -r '.versions[0].version')
# This jq expression should ensure that we always upload the
# highest-number version. Here is how this works:
#
# 1. The GitHub API documentation does not specify the order for the
# "list releases" endpoint, but does specify that the "latest"
# endpoint returns the release that points to the most recent commit.
# Assuming the same sort order is applied for the list endpoint
# (which empirically seems to hold so far), this means that they may
# be out-of-order wrt version numbers, e.g. 1.1.0 may appear after
# 1.0.2.
# 2. The `.tag_name | .[1:] | split (".") | map(tonumber)` part will
# turn "v1.0.2" into an array [1, 0, 2].
# 3. jq documents its sort method to sort numbers in numeric order
# and arrays in lexical order (ascending in both cases).
#
# This is required because, while the VSCode Marketplace does show
# _a_ version number, it doesn't handle versions at all: we can only
# have one version on the marketplace at any given time, and any
# upload replaces the existing version.
GITHUB=$(curl https://api.github.com/repos/digital-asset/daml/releases -s \
| jq -r '. | map(select(.prerelease == false)
| .tag_name
| .[1:]
| split (".")
| map(tonumber))
| sort
| reverse
| .[0]
| map(tostring)
| join(".")')
if [[ "$GITHUB" != "$MARKET" ]] && git merge-base --is-ancestor 798e96c9b9034eac85ace786b9e1955cf380285c v$GITHUB; then
echo "Publishing $GITHUB to VSCode Marketplace"
git checkout v$GITHUB
cd compiler/daml-extension
# This produces out/src/extension.js
bazel run @nodejs//:yarn
bazel run @nodejs//:yarn compile
bazel run --run_under="cd $PWD && " @daml_extension_deps//vsce/bin:vsce -- publish $GITHUB -p $MARKETPLACE_TOKEN
else
if [[ "$GITHUB" == "$MARKET" ]]; then
echo "Version on marketplace is already the latest ($GITHUB)."
else
echo "Latest version is not ready for marketplace publication."
fi
fi
env:
MARKETPLACE_TOKEN: $(VSCODE_MARKETPLACE_TOKEN)
- template: ci/tell-slack-failed.yml
- job: download_stats
timeoutInMinutes: 10
pool:
name: "linux-pool"
demands: assignment -equals default
steps:
- checkout: self
- bash: |
set -euo pipefail
eval "$(dev-env/bin/dade-assist)"
STATS=$(mktemp)
curl https://api.github.com/repos/digital-asset/daml/releases -s | gzip -9 > $STATS
GCS_KEY=$(mktemp)
cleanup () {
rm -f $GCS_KEY
}
trap cleanup EXIT
echo "$GOOGLE_APPLICATION_CREDENTIALS_CONTENT" > $GCS_KEY
gcloud auth activate-service-account --key-file=$GCS_KEY
BOTO_CONFIG=/dev/null gsutil cp $STATS gs://daml-data/downloads/$(date -u +%Y%m%d_%H%M%SZ).json.gz
env:
GOOGLE_APPLICATION_CREDENTIALS_CONTENT: $(GOOGLE_APPLICATION_CREDENTIALS_CONTENT)
- template: ci/tell-slack-failed.yml