awx/awx/main/scheduler/__init__.py

#Copyright (c) 2015 Ansible, Inc.
# All Rights Reserved

# Python
from datetime import timedelta
import logging
from sets import Set

# Django
from django.conf import settings
from django.db import transaction, connection
from django.db.utils import DatabaseError
from django.utils.translation import ugettext_lazy as _

# AWX
from awx.main.models import * # noqa
#from awx.main.scheduler.dag_simple import SimpleDAG
from awx.main.scheduler.dag_workflow import WorkflowDAG

from awx.main.scheduler.dependency_graph import DependencyGraph
from awx.main.scheduler.partial import (
    JobDict,
    ProjectUpdateDict,
    ProjectUpdateLatestDict,
    InventoryUpdateDict,
    InventoryUpdateLatestDict,
    InventorySourceDict,
    SystemJobDict,
    AdHocCommandDict,
    WorkflowJobDict,
)
from awx.main.tasks import _send_notification_templates

# Celery
from celery.task.control import inspect

logger = logging.getLogger('awx.main.scheduler')


class TaskManager():
    def __init__(self):
        self.graph = DependencyGraph()
        self.capacity_total = Instance.objects.total_capacity()
        self.capacity_used = 0

    def get_tasks(self):
        status_list = ('pending', 'waiting', 'running')

        jobs = JobDict.filter_partial(status=status_list)
        inventory_updates = InventoryUpdateDict.filter_partial(status=status_list)
        project_updates = ProjectUpdateDict.filter_partial(status=status_list)
        system_jobs = SystemJobDict.filter_partial(status=status_list)
        ad_hoc_commands = AdHocCommandDict.filter_partial(status=status_list)
        workflow_jobs = WorkflowJobDict.filter_partial(status=status_list)

        all_actions = sorted(jobs + project_updates + inventory_updates + system_jobs + ad_hoc_commands + workflow_jobs,
                             key=lambda task: task['created'])
        return all_actions

    '''
    Tasks that are running and SHOULD have a celery task.
    '''
    def get_running_tasks(self):
        status_list = ('running',)

        jobs = JobDict.filter_partial(status=status_list)
        inventory_updates = InventoryUpdateDict.filter_partial(status=status_list)
        project_updates = ProjectUpdateDict.filter_partial(status=status_list)
        system_jobs = SystemJobDict.filter_partial(status=status_list)
        ad_hoc_commands = AdHocCommandDict.filter_partial(status=status_list)

        all_actions = sorted(jobs + project_updates + inventory_updates + system_jobs + ad_hoc_commands,
                             key=lambda task: task['created'])
        return all_actions

    # TODO: Consider a database query for this logic
    def get_latest_project_update_tasks(self, all_sorted_tasks):
        project_ids = Set()
        for task in all_sorted_tasks:
            if type(task) == JobDict:
                project_ids.add(task['project_id'])

        return ProjectUpdateLatestDict.filter_partial(list(project_ids))

    # TODO: Consider a database query for this logic
    def get_latest_inventory_update_tasks(self, all_sorted_tasks):
        inventory_ids = Set()
        for task in all_sorted_tasks:
            if type(task) == JobDict:
                inventory_ids.add(task['inventory_id'])

        return InventoryUpdateLatestDict.filter_partial(list(inventory_ids))


    def get_running_workflow_jobs(self):
        graph_workflow_jobs = [wf for wf in
                               WorkflowJob.objects.filter(status='running')]
        return graph_workflow_jobs

    # TODO: Consider a database query for this logic
    def get_inventory_source_tasks(self, all_sorted_tasks):
        inventory_ids = Set()
        results = []
        for task in all_sorted_tasks:
            if type(task) is JobDict:
                inventory_ids.add(task['inventory_id'])

        for inventory_id in inventory_ids:
            results.append((inventory_id, InventorySourceDict.filter_partial(inventory_id)))

        return results

    def spawn_workflow_graph_jobs(self, workflow_jobs):
        for workflow_job in workflow_jobs:
            dag = WorkflowDAG(workflow_job)
            spawn_nodes = dag.bfs_nodes_to_run()
            for spawn_node in spawn_nodes:
                if spawn_node.unified_job_template is None:
                    continue
                kv = spawn_node.get_job_kwargs()
                job = spawn_node.unified_job_template.create_unified_job(**kv)
                spawn_node.job = job
                spawn_node.save()
                if job._resources_sufficient_for_launch():
                    can_start = job.signal_start(**kv)
                    if not can_start:
                        job.job_explanation = _("Job spawned from workflow could not start because it "
                                                "was not in the right state or required manual credentials")
                else:
                    can_start = False
                    job.job_explanation = _("Job spawned from workflow could not start because it "
                                            "was missing a related resource such as project or inventory")
                if not can_start:
                    job.status = 'failed'
                    job.save(update_fields=['status', 'job_explanation'])
                    connection.on_commit(lambda: job.websocket_emit_status('failed'))

                # TODO: should we emit a status on the socket here similar to tasks.py tower_periodic_scheduler() ?
                #emit_websocket_notification('/socket.io/jobs', '', dict(id=))

    # See comment in tasks.py::RunWorkflowJob::run()
    def process_finished_workflow_jobs(self, workflow_jobs):
        result = []
        for workflow_job in workflow_jobs:
            dag = WorkflowDAG(workflow_job)
            if workflow_job.cancel_flag:
                workflow_job.status = 'canceled'
                workflow_job.save()
                dag.cancel_node_jobs()
                connection.on_commit(lambda: workflow_job.websocket_emit_status(workflow_job.status))
            elif dag.is_workflow_done():
                result.append(workflow_job.id)
                if workflow_job._has_failed():
                    workflow_job.status = 'failed'
                else:
                    workflow_job.status = 'successful'
                workflow_job.save()
                connection.on_commit(lambda: workflow_job.websocket_emit_status(workflow_job.status))
        return result

    def get_active_tasks(self):
        inspector = inspect()
        if not hasattr(settings, 'IGNORE_CELERY_INSPECTOR'):
            active_task_queues = inspector.active()
        else:
            logger.warn("Ignoring celery task inspector")
            active_task_queues = None

        active_tasks = set()
        if active_task_queues is not None:
            for queue in active_task_queues:
                map(lambda at: active_tasks.add(at['id']), active_task_queues[queue])
        else:
            if not hasattr(settings, 'CELERY_UNIT_TEST'):
                return (None, None)

        return (active_task_queues, active_tasks)

    def get_dependent_jobs_for_inv_and_proj_update(self, job_obj):
        return [{'type': j.model_to_str(), 'id': j.id} for j in job_obj.dependent_jobs.all()]

    def start_task(self, task, dependent_tasks=[]):
        from awx.main.tasks import handle_work_error, handle_work_success

        task_actual = {
            'type':task.get_job_type_str(),
            'id': task['id'],
        }
        dependencies = [{'type': t.get_job_type_str(), 'id': t['id']} for t in dependent_tasks]

        error_handler = handle_work_error.s(subtasks=[task_actual] + dependencies)
        success_handler = handle_work_success.s(task_actual=task_actual)

        job_obj = task.get_full()
        '''
        This is to account for when there isn't enough capacity to execute all
        dependent jobs (i.e. proj or inv update) within the same schedule() 
        call.
        
        Proceeding calls to schedule() need to recontruct the proj or inv
        update -> job fail logic dependency. The below call recontructs that
        failure dependency.
        '''
        if len(dependencies) == 0:
            dependencies = self.get_dependent_jobs_for_inv_and_proj_update(job_obj)
        job_obj.status = 'waiting'

        (start_status, opts) = job_obj.pre_start()
        if not start_status:
            job_obj.status = 'failed'
            if job_obj.job_explanation:
                job_obj.job_explanation += ' '
            job_obj.job_explanation += 'Task failed pre-start check.'
            job_obj.save()
            # TODO: run error handler to fail sub-tasks and send notifications
        else:
            if type(job_obj) is WorkflowJob:
                job_obj.status = 'running'

            job_obj.save()

            self.consume_capacity(task)

        def post_commit():
            job_obj.websocket_emit_status(job_obj.status)
            if job_obj.status != 'failed':
                job_obj.start_celery_task(opts, error_callback=error_handler, success_callback=success_handler)

        connection.on_commit(post_commit)

    def process_runnable_tasks(self, runnable_tasks):
        map(lambda task: self.graph.add_job(task), runnable_tasks)

    def create_project_update(self, task):
        dep = Project.objects.get(id=task['project_id']).create_project_update(launch_type='dependency')

        # Project created 1 seconds behind
        dep.created = task['created'] - timedelta(seconds=1)
        dep.status = 'pending'
        dep.save()

        project_task = ProjectUpdateDict.get_partial(dep.id)

        return project_task

    def create_inventory_update(self, task, inventory_source_task):
        dep = InventorySource.objects.get(id=inventory_source_task['id']).create_inventory_update(launch_type='dependency')

        dep.created = task['created'] - timedelta(seconds=2)
        dep.status = 'pending'
        dep.save()

        inventory_task = InventoryUpdateDict.get_partial(dep.id)
        
        '''
        Update internal datastructures with the newly created inventory update
        '''
        # Should be only 1 inventory update. The one for the job (task)
        latest_inventory_updates = self.get_latest_inventory_update_tasks([task])
        self.process_latest_inventory_updates(latest_inventory_updates)

        inventory_sources = self.get_inventory_source_tasks([task])
        self.process_inventory_sources(inventory_sources)

        self.graph.add_job(inventory_task)

        return inventory_task

    '''
    Since we are dealing with partial objects we don't get to take advantage
    of Django to resolve the type of related Many to Many field dependent_job.

    Hence the, potentional, double query in this method.
    '''
    def get_related_dependent_jobs_as_patials(self, job_ids):
        dependent_partial_jobs = []
        for id in job_ids:
            if ProjectUpdate.objects.filter(id=id).exists():
                dependent_partial_jobs.append(ProjectUpdateDict({"id": id}).refresh_partial())
            elif InventoryUpdate.objects.filter(id=id).exists():
                dependent_partial_jobs.append(InventoryUpdateDict({"id": id}).refresh_partial())
        return dependent_partial_jobs

    def capture_chain_failure_dependencies(self, task, dependencies):
        for dep in dependencies:
            dep_obj = dep.get_full()
            dep_obj.dependent_jobs.add(task['id'])
            dep_obj.save()
            '''
            if not 'dependent_jobs__id' in task.data:
                task.data['dependent_jobs__id'] = [dep_obj.data['id']]
            else:
                task.data['dependent_jobs__id'].append(dep_obj.data['id'])
            '''

    def generate_dependencies(self, task):
        dependencies = []
        # TODO: What if the project is null ?
        if type(task) is JobDict:

            if task['project__scm_update_on_launch'] is True and \
                    self.graph.should_update_related_project(task):
                project_task = self.create_project_update(task)
                dependencies.append(project_task)
                # Inventory created 2 seconds behind job

            '''
            Inventory may have already been synced from a provision callback.
            '''
            inventory_sources_already_updated = task.get_inventory_sources_already_updated()

            '''
            get_inventory_sources() only return update on launch sources
            '''
            for inventory_source_task in self.graph.get_inventory_sources(task['inventory_id']):
                if inventory_source_task['id'] in inventory_sources_already_updated:
                    continue
                if self.graph.should_update_related_inventory_source(task, inventory_source_task['id']):
                    inventory_task = self.create_inventory_update(task, inventory_source_task)
                    dependencies.append(inventory_task)

        self.capture_chain_failure_dependencies(task, dependencies)
        return dependencies

    def process_latest_project_updates(self, latest_project_updates):
        map(lambda task: self.graph.add_latest_project_update(task), latest_project_updates)

    def process_latest_inventory_updates(self, latest_inventory_updates):
        map(lambda task: self.graph.add_latest_inventory_update(task), latest_inventory_updates)

    def process_inventory_sources(self, inventory_id_sources):
        map(lambda (inventory_id, inventory_sources): self.graph.add_inventory_sources(inventory_id, inventory_sources), inventory_id_sources)

    def process_dependencies(self, dependent_task, dependency_tasks):
        for task in dependency_tasks:
            # ProjectUpdate or InventoryUpdate may be blocked by another of
            # the same type.
            if not self.graph.is_job_blocked(task):
                self.graph.add_job(task)
                if not self.would_exceed_capacity(task):
                    self.start_task(task, [dependent_task])
            else:
                self.graph.add_job(task)

    def process_pending_tasks(self, pending_tasks):
        for task in pending_tasks:
            # Stop processing tasks if we know we are out of capacity
            if self.get_remaining_capacity() <= 0:
                return

            if not self.graph.is_job_blocked(task):
                dependencies = self.generate_dependencies(task)
                self.process_dependencies(task, dependencies)

                # Spawning deps might have blocked us
                if not self.graph.is_job_blocked(task):
                    self.graph.add_job(task)
                    if not self.would_exceed_capacity(task):
                        self.start_task(task)
                else:
                    self.graph.add_job(task)

    def process_celery_tasks(self, active_tasks, all_running_sorted_tasks):
        '''
        Rectify tower db <-> celery inconsistent view of jobs state
        '''
        for task in all_running_sorted_tasks:

            if (task['celery_task_id'] not in active_tasks and not hasattr(settings, 'IGNORE_CELERY_INSPECTOR')):
                # TODO: try catch the getting of the job. The job COULD have been deleted
                task_obj = task.get_full()
                # Ensure job did not finish running between the time we get the
                # list of task id's from celery and now.
                # Note: This is an actual fix, not a reduction in the time 
                # window that this can happen.
                if task_obj.status != 'running':
                    continue
                task_obj.status = 'failed'
                task_obj.job_explanation += ' '.join((
                    'Task was marked as running in Tower but was not present in',
                    'Celery, so it has been marked as failed.',
                ))
                task_obj.save()
                _send_notification_templates(task_obj, 'failed')
                task_obj.websocket_emit_status('failed')

                logger.error("Task %s appears orphaned... marking as failed" % task)


    def calculate_capacity_used(self, tasks):
        self.capacity_used = 0
        for t in tasks:
            self.capacity_used += t.task_impact()

    def would_exceed_capacity(self, task):
        if self.capacity_used == 0:
            return False
        return (task.task_impact() + self.capacity_used > self.capacity_total)

    def consume_capacity(self, task):
        self.capacity_used += task.task_impact()

    def get_remaining_capacity(self):
        return (self.capacity_total - self.capacity_used)

    def process_tasks(self, all_sorted_tasks):
        running_tasks = filter(lambda t: t['status'] == 'running', all_sorted_tasks)
        runnable_tasks = filter(lambda t: t['status'] in ['waiting', 'running'], all_sorted_tasks)

        self.calculate_capacity_used(running_tasks)

        self.process_runnable_tasks(runnable_tasks)

        pending_tasks = filter(lambda t: t['status'] in 'pending', all_sorted_tasks)
        self.process_pending_tasks(pending_tasks)

    def _schedule(self):
        finished_wfjs = []
        all_sorted_tasks = self.get_tasks()
        if len(all_sorted_tasks) > 0:
            latest_project_updates = self.get_latest_project_update_tasks(all_sorted_tasks)
            self.process_latest_project_updates(latest_project_updates)

            latest_inventory_updates = self.get_latest_inventory_update_tasks(all_sorted_tasks)
            self.process_latest_inventory_updates(latest_inventory_updates)

            inventory_id_sources = self.get_inventory_source_tasks(all_sorted_tasks)
            self.process_inventory_sources(inventory_id_sources)

            running_workflow_tasks = self.get_running_workflow_jobs()
            finished_wfjs = self.process_finished_workflow_jobs(running_workflow_tasks)

            self.spawn_workflow_graph_jobs(running_workflow_tasks)

            self.process_tasks(all_sorted_tasks)
        return finished_wfjs

    def schedule(self):
        with transaction.atomic():
            # Lock
            try:
                Instance.objects.select_for_update(nowait=True).all()[0]
            except DatabaseError:
                return

            finished_wfjs = self._schedule()

        # Operations whose queries rely on modifications made during the atomic scheduling session
        for wfj in WorkflowJob.objects.filter(id__in=finished_wfjs):
            _send_notification_templates(wfj, 'succeeded' if wfj.status == 'successful' else 'failed')