fune/taskcluster/gecko_taskgraph/decision.py

538 lines
18 KiB
Python

# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this
# file, You can obtain one at http://mozilla.org/MPL/2.0/.
import os
import json
import logging
import time
import shutil
import sys
from collections import defaultdict
import yaml
from redo import retry
from taskgraph.parameters import Parameters
from taskgraph.taskgraph import TaskGraph
from taskgraph.util.taskcluster import get_artifact
from taskgraph.util.yaml import load_yaml
from voluptuous import Required, Optional, Any
from . import GECKO, create
from .actions import render_actions_json
from .create import create_tasks
from .generator import TaskGraphGenerator
from .parameters import get_version, get_app_version
from .try_option_syntax import parse_message
from .util.backstop import is_backstop, BACKSTOP_INDEX
from .util.bugbug import push_schedules
from .util.chunking import resolver
from .util.hg import get_hg_revision_branch, get_hg_commit_message
from .util.partials import populate_release_history
from .util.python_path import find_object
from .util.schema import validate_schema, Schema
from .util.taskcluster import insert_index
from .util.taskgraph import find_decision_task, find_existing_tasks_from_previous_kinds
logger = logging.getLogger(__name__)
ARTIFACTS_DIR = "artifacts"
# For each project, this gives a set of parameters specific to the project.
# See `taskcluster/docs/parameters.rst` for information on parameters.
PER_PROJECT_PARAMETERS = {
"try": {
"target_tasks_method": "try_tasks",
},
"kaios-try": {
"target_tasks_method": "try_tasks",
},
"ash": {
"target_tasks_method": "default",
},
"cedar": {
"target_tasks_method": "default",
},
"oak": {
"target_tasks_method": "nightly_desktop",
"release_type": "nightly-oak",
},
"graphics": {
"target_tasks_method": "graphics_tasks",
},
"autoland": {
"optimize_strategies": "gecko_taskgraph.optimize:project.autoland",
"target_tasks_method": "autoland_tasks",
"test_manifest_loader": "bugbug", # Remove this line to disable "manifest scheduling".
},
"mozilla-central": {
"target_tasks_method": "mozilla_central_tasks",
"release_type": "nightly",
},
"mozilla-beta": {
"target_tasks_method": "mozilla_beta_tasks",
"release_type": "beta",
},
"mozilla-release": {
"target_tasks_method": "mozilla_release_tasks",
"release_type": "release",
},
"mozilla-esr91": {
"target_tasks_method": "mozilla_esr91_tasks",
"release_type": "esr91",
},
"mozilla-esr102": {
"target_tasks_method": "mozilla_esr102_tasks",
"release_type": "esr102",
},
"pine": {
"target_tasks_method": "pine_tasks",
},
"kaios": {
"target_tasks_method": "kaios_tasks",
},
# the default parameters are used for projects that do not match above.
"default": {
"target_tasks_method": "default",
},
}
try_task_config_schema = Schema(
{
Required("tasks"): [str],
Optional("browsertime"): bool,
Optional("chemspill-prio"): bool,
Optional("disable-pgo"): bool,
Optional("env"): {str: str},
Optional("gecko-profile"): bool,
Optional("gecko-profile-interval"): float,
Optional("gecko-profile-entries"): int,
Optional("gecko-profile-features"): str,
Optional("gecko-profile-threads"): str,
Optional(
"perftest-options",
description="Options passed from `mach perftest` to try.",
): object,
Optional(
"optimize-strategies",
description="Alternative optimization strategies to use instead of the default. "
"A module path pointing to a dict to be use as the `strategy_override` "
"argument in `gecko_taskgraph.optimize.optimize_task_graph`.",
): str,
Optional("rebuild"): int,
Optional("tasks-regex"): {
"include": Any(None, [str]),
"exclude": Any(None, [str]),
},
Optional("use-artifact-builds"): bool,
Optional(
"worker-overrides",
description="Mapping of worker alias to worker pools to use for those aliases.",
): {str: str},
Optional("routes"): [str],
}
)
"""
Schema for try_task_config.json files.
"""
try_task_config_schema_v2 = Schema(
{
Optional("parameters"): {str: object},
}
)
def full_task_graph_to_runnable_jobs(full_task_json):
runnable_jobs = {}
for label, node in full_task_json.items():
if not ("extra" in node["task"] and "treeherder" in node["task"]["extra"]):
continue
th = node["task"]["extra"]["treeherder"]
runnable_jobs[label] = {"symbol": th["symbol"]}
for i in ("groupName", "groupSymbol", "collection"):
if i in th:
runnable_jobs[label][i] = th[i]
if th.get("machine", {}).get("platform"):
runnable_jobs[label]["platform"] = th["machine"]["platform"]
return runnable_jobs
def full_task_graph_to_manifests_by_task(full_task_json):
manifests_by_task = defaultdict(list)
for label, node in full_task_json.items():
manifests = node["attributes"].get("test_manifests")
if not manifests:
continue
manifests_by_task[label].extend(manifests)
return manifests_by_task
def try_syntax_from_message(message):
"""
Parse the try syntax out of a commit message, returning '' if none is
found.
"""
try_idx = message.find("try:")
if try_idx == -1:
return ""
return message[try_idx:].split("\n", 1)[0]
def taskgraph_decision(options, parameters=None):
"""
Run the decision task. This function implements `mach taskgraph decision`,
and is responsible for
* processing decision task command-line options into parameters
* running task-graph generation exactly the same way the other `mach
taskgraph` commands do
* generating a set of artifacts to memorialize the graph
* calling TaskCluster APIs to create the graph
"""
parameters = parameters or (
lambda graph_config: get_decision_parameters(graph_config, options)
)
decision_task_id = os.environ["TASK_ID"]
# create a TaskGraphGenerator instance
tgg = TaskGraphGenerator(
root_dir=options.get("root"),
parameters=parameters,
decision_task_id=decision_task_id,
write_artifacts=True,
)
if not create.testing:
# set additional index paths for the decision task
set_decision_indexes(decision_task_id, tgg.parameters, tgg.graph_config)
# write out the parameters used to generate this graph
write_artifact("parameters.yml", dict(**tgg.parameters))
# write out the public/actions.json file
write_artifact(
"actions.json",
render_actions_json(tgg.parameters, tgg.graph_config, decision_task_id),
)
# write out the full graph for reference
full_task_json = tgg.full_task_graph.to_json()
write_artifact("full-task-graph.json", full_task_json)
# write out the public/runnable-jobs.json file
write_artifact(
"runnable-jobs.json", full_task_graph_to_runnable_jobs(full_task_json)
)
# write out the public/manifests-by-task.json file
write_artifact(
"manifests-by-task.json.gz",
full_task_graph_to_manifests_by_task(full_task_json),
)
# write out the public/tests-by-manifest.json file
write_artifact("tests-by-manifest.json.gz", resolver.tests_by_manifest)
# this is just a test to check whether the from_json() function is working
_, _ = TaskGraph.from_json(full_task_json)
# write out the target task set to allow reproducing this as input
write_artifact("target-tasks.json", list(tgg.target_task_set.tasks.keys()))
# write out the optimized task graph to describe what will actually happen,
# and the map of labels to taskids
write_artifact("task-graph.json", tgg.morphed_task_graph.to_json())
write_artifact("label-to-taskid.json", tgg.label_to_taskid)
# write bugbug scheduling information if it was invoked
if len(push_schedules) > 0:
write_artifact("bugbug-push-schedules.json", push_schedules.popitem()[1])
# cache run-task & misc/fetch-content
scripts_root_dir = os.path.join(
"/builds/worker/checkouts/gecko/taskcluster/scripts"
)
run_task_file_path = os.path.join(scripts_root_dir, "run-task")
fetch_content_file_path = os.path.join(scripts_root_dir, "misc/fetch-content")
shutil.copy2(run_task_file_path, ARTIFACTS_DIR)
shutil.copy2(fetch_content_file_path, ARTIFACTS_DIR)
# actually create the graph
create_tasks(
tgg.graph_config,
tgg.morphed_task_graph,
tgg.label_to_taskid,
tgg.parameters,
decision_task_id=decision_task_id,
)
def get_decision_parameters(graph_config, options):
"""
Load parameters from the command-line options for 'taskgraph decision'.
This also applies per-project parameters, based on the given project.
"""
product_dir = graph_config["product-dir"]
parameters = {
n: options[n]
for n in [
"base_repository",
"head_repository",
"head_rev",
"head_ref",
"head_tag",
"project",
"pushlog_id",
"pushdate",
"owner",
"level",
"repository_type",
"target_tasks_method",
"tasks_for",
]
if n in options
}
for n in (
"comm_base_repository",
"comm_head_repository",
"comm_head_rev",
"comm_head_ref",
):
if n in options and options[n] is not None:
parameters[n] = options[n]
commit_message = get_hg_commit_message(os.path.join(GECKO, product_dir))
# Define default filter list, as most configurations shouldn't need
# custom filters.
parameters["filters"] = [
"target_tasks_method",
]
parameters["existing_tasks"] = {}
parameters["do_not_optimize"] = []
parameters["build_number"] = 1
parameters["version"] = get_version(product_dir)
parameters["app_version"] = get_app_version(product_dir)
parameters["message"] = try_syntax_from_message(commit_message)
parameters["hg_branch"] = get_hg_revision_branch(
GECKO, revision=parameters["head_rev"]
)
parameters["next_version"] = None
parameters["optimize_strategies"] = None
parameters["optimize_target_tasks"] = True
parameters["phabricator_diff"] = None
parameters["release_type"] = ""
parameters["release_eta"] = ""
parameters["release_enable_partner_repack"] = False
parameters["release_enable_partner_attribution"] = False
parameters["release_partners"] = []
parameters["release_partner_config"] = {}
parameters["release_partner_build_number"] = 1
parameters["release_enable_emefree"] = False
parameters["release_product"] = None
parameters["required_signoffs"] = []
parameters["signoff_urls"] = {}
parameters["test_manifest_loader"] = "default"
parameters["try_mode"] = None
parameters["try_task_config"] = {}
parameters["try_options"] = None
# owner must be an email, but sometimes (e.g., for ffxbld) it is not, in which
# case, fake it
if "@" not in parameters["owner"]:
parameters["owner"] += "@noreply.mozilla.org"
# use the pushdate as build_date if given, else use current time
parameters["build_date"] = parameters["pushdate"] or int(time.time())
# moz_build_date is the build identifier based on build_date
parameters["moz_build_date"] = time.strftime(
"%Y%m%d%H%M%S", time.gmtime(parameters["build_date"])
)
project = parameters["project"]
try:
parameters.update(PER_PROJECT_PARAMETERS[project])
except KeyError:
logger.warning(
"using default project parameters; add {} to "
"PER_PROJECT_PARAMETERS in {} to customize behavior "
"for this project".format(project, __file__)
)
parameters.update(PER_PROJECT_PARAMETERS["default"])
# `target_tasks_method` has higher precedence than `project` parameters
if options.get("target_tasks_method"):
parameters["target_tasks_method"] = options["target_tasks_method"]
# ..but can be overridden by the commit message: if it contains the special
# string "DONTBUILD" and this is an on-push decision task, then use the
# special 'nothing' target task method.
if "DONTBUILD" in commit_message and options["tasks_for"] == "hg-push":
parameters["target_tasks_method"] = "nothing"
if options.get("include_push_tasks"):
get_existing_tasks(options.get("rebuild_kinds", []), parameters, graph_config)
# If the target method is nightly, we should build partials. This means
# knowing what has been released previously.
# An empty release_history is fine, it just means no partials will be built
parameters.setdefault("release_history", dict())
if "nightly" in parameters.get("target_tasks_method", ""):
parameters["release_history"] = populate_release_history("Firefox", project)
if options.get("try_task_config_file"):
task_config_file = os.path.abspath(options.get("try_task_config_file"))
else:
# if try_task_config.json is present, load it
task_config_file = os.path.join(os.getcwd(), "try_task_config.json")
# load try settings
if "try" in project and options["tasks_for"] == "hg-push":
set_try_config(parameters, task_config_file)
if options.get("optimize_target_tasks") is not None:
parameters["optimize_target_tasks"] = options["optimize_target_tasks"]
# Determine if this should be a backstop push.
parameters["backstop"] = is_backstop(parameters)
if "decision-parameters" in graph_config["taskgraph"]:
find_object(graph_config["taskgraph"]["decision-parameters"])(
graph_config, parameters
)
result = Parameters(**parameters)
result.check()
return result
def get_existing_tasks(rebuild_kinds, parameters, graph_config):
"""
Find the decision task corresponding to the on-push graph, and return
a mapping of labels to task-ids from it. This will skip the kinds specificed
by `rebuild_kinds`.
"""
try:
decision_task = retry(
find_decision_task,
args=(parameters, graph_config),
attempts=4,
sleeptime=5 * 60,
)
except Exception:
logger.exception("Didn't find existing push task.")
sys.exit(1)
_, task_graph = TaskGraph.from_json(
get_artifact(decision_task, "public/full-task-graph.json")
)
parameters["existing_tasks"] = find_existing_tasks_from_previous_kinds(
task_graph, [decision_task], rebuild_kinds
)
def set_try_config(parameters, task_config_file):
if os.path.isfile(task_config_file):
logger.info(f"using try tasks from {task_config_file}")
with open(task_config_file) as fh:
task_config = json.load(fh)
task_config_version = task_config.pop("version", 1)
if task_config_version == 1:
validate_schema(
try_task_config_schema,
task_config,
"Invalid v1 `try_task_config.json`.",
)
parameters["try_mode"] = "try_task_config"
parameters["try_task_config"] = task_config
elif task_config_version == 2:
validate_schema(
try_task_config_schema_v2,
task_config,
"Invalid v2 `try_task_config.json`.",
)
parameters.update(task_config["parameters"])
return
else:
raise Exception(
f"Unknown `try_task_config.json` version: {task_config_version}"
)
if "try:" in parameters["message"]:
parameters["try_mode"] = "try_option_syntax"
parameters.update(parse_message(parameters["message"]))
else:
parameters["try_options"] = None
if parameters["try_mode"] == "try_task_config":
# The user has explicitly requested a set of jobs, so run them all
# regardless of optimization. Their dependencies can be optimized,
# though.
parameters["optimize_target_tasks"] = False
else:
# For a try push with no task selection, apply the default optimization
# process to all of the tasks.
parameters["optimize_target_tasks"] = True
def set_decision_indexes(decision_task_id, params, graph_config):
index_paths = []
if params["backstop"]:
index_paths.append(BACKSTOP_INDEX)
subs = params.copy()
subs["trust-domain"] = graph_config["trust-domain"]
index_paths = [i.format(**subs) for i in index_paths]
for index_path in index_paths:
insert_index(index_path, decision_task_id, use_proxy=True)
def write_artifact(filename, data):
logger.info(f"writing artifact file `{filename}`")
if not os.path.isdir(ARTIFACTS_DIR):
os.mkdir(ARTIFACTS_DIR)
path = os.path.join(ARTIFACTS_DIR, filename)
if filename.endswith(".yml"):
with open(path, "w") as f:
yaml.safe_dump(data, f, allow_unicode=True, default_flow_style=False)
elif filename.endswith(".json"):
with open(path, "w") as f:
json.dump(data, f, sort_keys=True, indent=2, separators=(",", ": "))
elif filename.endswith(".json.gz"):
import gzip
with gzip.open(path, "wb") as f:
f.write(json.dumps(data).encode("utf-8"))
else:
raise TypeError(f"Don't know how to write to {filename}")
def read_artifact(filename):
path = os.path.join(ARTIFACTS_DIR, filename)
if filename.endswith(".yml"):
return load_yaml(path, filename)
elif filename.endswith(".json"):
with open(path) as f:
return json.load(f)
elif filename.endswith(".json.gz"):
import gzip
with gzip.open(path, "rb") as f:
return json.load(f.decode("utf-8"))
else:
raise TypeError(f"Don't know how to read {filename}")
def rename_artifact(src, dest):
os.rename(os.path.join(ARTIFACTS_DIR, src), os.path.join(ARTIFACTS_DIR, dest))