From 9d06a61fe4546c9ddae550170f74bd483fad7269 Mon Sep 17 00:00:00 2001 From: Renat Akhmerov Date: Tue, 13 Sep 2016 12:37:20 +0300 Subject: [PATCH] Optimize the logic that check if 'join' task is allowed to start * Moved DB related lookup functions from workflow/utils to a separate module lookup_utils * Optimized data access pattern where we calculate 'join' states induced by upstream tasks, applied caching for some lookup operations * Added caching in direct workflow specification for inbound and outbound task specs, if workflow size is large calculating them may be expensive * Added an adaptive delay between calls that refresh 'join' task state based on a number of unfulfilled preconditions returned by workflow controller Change-Id: I383fa52f2f05877df7522048020cc7ff280324a2 --- mistral/engine/task_handler.py | 20 ++++- mistral/engine/workflows.py | 10 ++- mistral/tests/unit/base.py | 3 + mistral/utils/yaql_utils.py | 12 +-- mistral/workbook/v2/workflows.py | 47 +++++++++- mistral/workflow/base.py | 20 +++-- mistral/workflow/direct_workflow.py | 129 +++++++++++++++++---------- mistral/workflow/lookup_utils.py | 109 ++++++++++++++++++++++ mistral/workflow/reverse_workflow.py | 20 +++-- mistral/workflow/utils.py | 45 ---------- 10 files changed, 293 insertions(+), 122 deletions(-) create mode 100644 mistral/workflow/lookup_utils.py diff --git a/mistral/engine/task_handler.py b/mistral/engine/task_handler.py index 4a630f46a..0c1e7bfb4 100644 --- a/mistral/engine/task_handler.py +++ b/mistral/engine/task_handler.py @@ -73,7 +73,7 @@ def run_task(wf_cmd): return if task.is_waiting() and (task.is_created() or task.is_state_changed()): - _schedule_refresh_task_state(task.task_ex) + _schedule_refresh_task_state(task.task_ex, 1) @profiler.trace('task-handler-on-action-complete') @@ -251,15 +251,27 @@ def _refresh_task_state(task_ex_id): wf_spec ) - state, state_info = wf_ctrl.get_logical_task_state(task_ex) + state, state_info, cardinality = wf_ctrl.get_logical_task_state( + task_ex + ) if state == states.RUNNING: continue_task(task_ex) elif state == states.ERROR: fail_task(task_ex, state_info) elif state == states.WAITING: - # TODO(rakhmerov): Algorithm for increasing rescheduling delay. - _schedule_refresh_task_state(task_ex, 1) + # Let's assume that a task takes 0.01 sec in average to complete + # and based on this assumption calculate a time of the next check. + # The estimation is very rough, of course, but this delay will be + # decreasing as task preconditions will be completing which will + # give a decent asymptotic approximation. + # For example, if a 'join' task has 100 inbound incomplete tasks + # then the next 'refresh_task_state' call will happen in 10 + # seconds. For 500 tasks it will be 50 seconds. The larger the + # workflow is, the more beneficial this mechanism will be. + delay = int(cardinality * 0.01) + + _schedule_refresh_task_state(task_ex, max(1, delay)) else: # Must never get here. raise RuntimeError( diff --git a/mistral/engine/workflows.py b/mistral/engine/workflows.py index 1cf874393..ae0ba044c 100644 --- a/mistral/engine/workflows.py +++ b/mistral/engine/workflows.py @@ -34,6 +34,7 @@ from mistral.workbook import parser as spec_parser from mistral.workflow import base as wf_base from mistral.workflow import commands from mistral.workflow import data_flow +from mistral.workflow import lookup_utils from mistral.workflow import states from mistral.workflow import utils as wf_utils @@ -158,6 +159,11 @@ class Workflow(object): assert self.wf_ex + # Since some lookup utils functions may use cache for completed tasks + # we need to clean caches to make sure that stale objects can't be + # retrieved. + lookup_utils.clean_caches() + wf_service.update_workflow_execution_env(self.wf_ex, env) self.set_state(states.RUNNING, recursive=True) @@ -429,7 +435,7 @@ def _build_fail_info_message(wf_ctrl, wf_ex): failed_tasks = sorted( filter( lambda t: not wf_ctrl.is_error_handled_for(t), - wf_utils.find_error_task_executions(wf_ex) + lookup_utils.find_error_task_executions(wf_ex.id) ), key=lambda t: t.name ) @@ -468,7 +474,7 @@ def _build_fail_info_message(wf_ctrl, wf_ex): def _build_cancel_info_message(wf_ctrl, wf_ex): # Try to find where cancel is exactly. cancelled_tasks = sorted( - wf_utils.find_cancelled_task_executions(wf_ex), + lookup_utils.find_cancelled_task_executions(wf_ex.id), key=lambda t: t.name ) diff --git a/mistral/tests/unit/base.py b/mistral/tests/unit/base.py index 497fc8305..8952d3bc9 100644 --- a/mistral/tests/unit/base.py +++ b/mistral/tests/unit/base.py @@ -36,6 +36,7 @@ from mistral.tests.unit import config as test_config from mistral.utils import inspect_utils as i_utils from mistral import version from mistral.workbook import parser as spec_parser +from mistral.workflow import lookup_utils RESOURCES_PATH = 'tests/resources/' LOG = logging.getLogger(__name__) @@ -244,6 +245,8 @@ class DbTestCase(BaseTest): action_manager.sync_db() def _clean_db(self): + lookup_utils.clean_caches() + contexts = [ get_context(default=False), get_context(default=True) diff --git a/mistral/utils/yaql_utils.py b/mistral/utils/yaql_utils.py index 122390fba..ba56b5487 100644 --- a/mistral/utils/yaql_utils.py +++ b/mistral/utils/yaql_utils.py @@ -13,13 +13,12 @@ # limitations under the License. +from oslo_serialization import jsonutils +from stevedore import extension import yaql from mistral.db.v2 import api as db_api from mistral import utils -from mistral.workflow import utils as wf_utils -from oslo_serialization import jsonutils -from stevedore import extension ROOT_CONTEXT = None @@ -87,8 +86,6 @@ def task_(context, task_name): # Importing data_flow in order to break cycle dependency between modules. from mistral.workflow import data_flow - wf_ex = db_api.get_workflow_execution(context['__execution']['id']) - # This section may not exist in a context if it's calculated not in # task scope. cur_task = context['__task_execution'] @@ -96,7 +93,10 @@ def task_(context, task_name): if cur_task and cur_task['name'] == task_name: task_ex = db_api.get_task_execution(cur_task['id']) else: - task_execs = wf_utils.find_task_executions_by_name(wf_ex, task_name) + task_execs = db_api.get_task_executions( + workflow_execution_id=context['__execution']['id'], + name=task_name + ) # TODO(rakhmerov): Account for multiple executions (i.e. in case of # cycles). diff --git a/mistral/workbook/v2/workflows.py b/mistral/workbook/v2/workflows.py index fe150cd36..16312983f 100644 --- a/mistral/workbook/v2/workflows.py +++ b/mistral/workbook/v2/workflows.py @@ -15,6 +15,7 @@ from oslo_utils import uuidutils import six +import threading from mistral import exceptions as exc from mistral import utils @@ -150,6 +151,18 @@ class DirectWorkflowSpec(WorkflowSpec): } } + def __init__(self, data): + super(DirectWorkflowSpec, self).__init__(data) + + # Init simple dictionary based caches for inbound and + # outbound task specifications. In fact, we don't need + # any special cache implementations here because these + # structures can't grow indefinitely. + self.inbound_tasks_cache_lock = threading.RLock() + self.inbound_tasks_cache = {} + self.outbound_tasks_cache_lock = threading.RLock() + self.outbound_tasks_cache = {} + def validate_semantics(self): super(DirectWorkflowSpec, self).validate_semantics() @@ -211,17 +224,43 @@ class DirectWorkflowSpec(WorkflowSpec): ] def find_inbound_task_specs(self, task_spec): - return [ + task_name = task_spec.get_name() + + with self.inbound_tasks_cache_lock: + specs = self.inbound_tasks_cache.get(task_name) + + if specs is not None: + return specs + + specs = [ t_s for t_s in self.get_tasks() - if self.transition_exists(t_s.get_name(), task_spec.get_name()) + if self.transition_exists(t_s.get_name(), task_name) ] + with self.inbound_tasks_cache_lock: + self.inbound_tasks_cache[task_name] = specs + + return specs + def find_outbound_task_specs(self, task_spec): - return [ + task_name = task_spec.get_name() + + with self.outbound_tasks_cache_lock: + specs = self.outbound_tasks_cache.get(task_name) + + if specs is not None: + return specs + + specs = [ t_s for t_s in self.get_tasks() - if self.transition_exists(task_spec.get_name(), t_s.get_name()) + if self.transition_exists(task_name, t_s.get_name()) ] + with self.outbound_tasks_cache_lock: + self.outbound_tasks_cache[task_name] = specs + + return specs + def has_inbound_transitions(self, task_spec): return len(self.find_inbound_task_specs(task_spec)) > 0 diff --git a/mistral/workflow/base.py b/mistral/workflow/base.py index 50fecb697..76f9d480e 100644 --- a/mistral/workflow/base.py +++ b/mistral/workflow/base.py @@ -26,13 +26,14 @@ from mistral import utils as u from mistral.workbook import parser as spec_parser from mistral.workflow import commands from mistral.workflow import data_flow +from mistral.workflow import lookup_utils from mistral.workflow import states -from mistral.workflow import utils as wf_utils LOG = logging.getLogger(__name__) +@profiler.trace('wf-controller-get-controller') def get_controller(wf_ex, wf_spec=None): """Gets a workflow controller instance by given workflow execution object. @@ -130,8 +131,13 @@ class WorkflowController(object): """Determines a logical state of the given task. :param task_ex: Task execution. - :return: Tuple (state, state_info) which the given task should have - according to workflow rules and current states of other tasks. + :return: Tuple (state, state_info, cardinality) where 'state' and + 'state_info' are the corresponding values which the given + task should have according to workflow rules and current + states of other tasks. 'cardinality' gives the estimation on + the number of preconditions that are not yet met in case if + state is WAITING. This number can be used to estimate how + frequently we can refresh the state of this task. """ raise NotImplementedError @@ -159,7 +165,9 @@ class WorkflowController(object): :return: True if there is one or more tasks in cancelled state. """ - return len(wf_utils.find_cancelled_task_executions(self.wf_ex)) > 0 + t_execs = lookup_utils.find_cancelled_task_executions(self.wf_ex.id) + + return len(t_execs) > 0 @abc.abstractmethod def evaluate_workflow_final_context(self): @@ -214,8 +222,8 @@ class WorkflowController(object): return [] # Add all tasks in IDLE state. - idle_tasks = wf_utils.find_task_executions_with_state( - self.wf_ex, + idle_tasks = lookup_utils.find_task_executions_with_state( + self.wf_ex.id, states.IDLE ) diff --git a/mistral/workflow/direct_workflow.py b/mistral/workflow/direct_workflow.py index 92092832c..17f294d92 100644 --- a/mistral/workflow/direct_workflow.py +++ b/mistral/workflow/direct_workflow.py @@ -13,6 +13,7 @@ # limitations under the License. from oslo_log import log as logging +from osprofiler import profiler from mistral import exceptions as exc from mistral import expressions as expr @@ -20,8 +21,8 @@ from mistral import utils from mistral.workflow import base from mistral.workflow import commands from mistral.workflow import data_flow +from mistral.workflow import lookup_utils from mistral.workflow import states -from mistral.workflow import utils as wf_utils LOG = logging.getLogger(__name__) @@ -46,8 +47,8 @@ class DirectWorkflowController(base.WorkflowController): return list( filter( lambda t_e: self._is_upstream_task_execution(task_spec, t_e), - wf_utils.find_task_executions_by_specs( - self.wf_ex, + lookup_utils.find_task_executions_by_specs( + self.wf_ex.id, self.wf_spec.find_inbound_task_specs(task_spec) ) ) @@ -60,7 +61,7 @@ class DirectWorkflowController(base.WorkflowController): if not t_spec.get_join(): return not t_ex_candidate.processed - induced_state = self._get_induced_join_state( + induced_state, _ = self._get_induced_join_state( self.wf_spec.get_tasks()[t_ex_candidate.name], t_spec ) @@ -173,7 +174,7 @@ class DirectWorkflowController(base.WorkflowController): # A simple 'non-join' task does not have any preconditions # based on state of other tasks so its logical state always # equals to its real state. - return task_ex.state, task_ex.state_info + return task_ex.state, task_ex.state_info, 0 return self._get_join_logical_state(task_spec) @@ -181,8 +182,7 @@ class DirectWorkflowController(base.WorkflowController): return bool(self.wf_spec.get_on_error_clause(task_ex.name)) def all_errors_handled(self): - for t_ex in wf_utils.find_error_task_executions(self.wf_ex): - + for t_ex in lookup_utils.find_error_task_executions(self.wf_ex.id): tasks_on_error = self._find_next_tasks_for_clause( self.wf_spec.get_on_error_clause(t_ex.name), data_flow.evaluate_task_outbound_context(t_ex) @@ -197,7 +197,7 @@ class DirectWorkflowController(base.WorkflowController): return list( filter( lambda t_ex: not self._has_outbound_tasks(t_ex), - wf_utils.find_successful_task_executions(self.wf_ex) + lookup_utils.find_successful_task_executions(self.wf_ex.id) ) ) @@ -270,64 +270,94 @@ class DirectWorkflowController(base.WorkflowController): if not condition or expr.evaluate(condition, ctx) ] + @profiler.trace('direct-wf-controller-get-join-logical-state') def _get_join_logical_state(self, task_spec): + """Evaluates logical state of 'join' task. + + :param task_spec: 'join' task specification. + :return: Tuple (state, state_info, spec_cardinality) where 'state' and + 'state_info' describe the logical state of the given 'join' + task and 'spec_cardinality' gives the remaining number of + unfulfilled preconditions. If logical state is not WAITING then + 'spec_cardinality' should always be 0. + """ + # TODO(rakhmerov): We need to use task_ex instead of task_spec # in order to cover a use case when there's more than one instance # of the same 'join' task in a workflow. + # TODO(rakhmerov): In some cases this method will be expensive because + # it uses a multistep recursive search. We need to optimize it moving + # forward (e.g. with Workflow Execution Graph). + join_expr = task_spec.get_join() in_task_specs = self.wf_spec.find_inbound_task_specs(task_spec) if not in_task_specs: - return states.RUNNING + return states.RUNNING, None, 0 - # List of tuples (task_name, state). + # List of tuples (task_name, (state, depth)). induced_states = [ (t_s.get_name(), self._get_induced_join_state(t_s, task_spec)) for t_s in in_task_specs ] def count(state): - return len(list(filter(lambda s: s[1] == state, induced_states))) + cnt = 0 + total_depth = 0 - error_count = count(states.ERROR) - running_count = count(states.RUNNING) + for s in induced_states: + if s[1][0] == state: + cnt += 1 + total_depth += s[1][1] + + return cnt, total_depth + + errors_tuples = count(states.ERROR) + runnings_tuple = count(states.RUNNING) total_count = len(induced_states) def _blocked_message(): return ( 'Blocked by tasks: %s' % - [s[0] for s in induced_states if s[1] == states.WAITING] + [s[0] for s in induced_states if s[1][0] == states.WAITING] ) def _failed_message(): return ( 'Failed by tasks: %s' % - [s[0] for s in induced_states if s[1] == states.ERROR] + [s[0] for s in induced_states if s[1][0] == states.ERROR] ) # If "join" is configured as a number or 'one'. if isinstance(join_expr, int) or join_expr == 'one': - cardinality = 1 if join_expr == 'one' else join_expr + spec_cardinality = 1 if join_expr == 'one' else join_expr - if running_count >= cardinality: - return states.RUNNING, None + if runnings_tuple[0] >= spec_cardinality: + return states.RUNNING, None, 0 # E.g. 'join: 3' with inbound [ERROR, ERROR, RUNNING, WAITING] # No chance to get 3 RUNNING states. - if error_count > (total_count - cardinality): - return states.ERROR, _failed_message() + if errors_tuples[0] > (total_count - spec_cardinality): + return states.ERROR, _failed_message(), 0 - return states.WAITING, _blocked_message() + # Calculate how many tasks need to finish to trigger this 'join'. + cardinality = spec_cardinality - runnings_tuple[0] + + return states.WAITING, _blocked_message(), cardinality if join_expr == 'all': - if total_count == running_count: - return states.RUNNING, None + if total_count == runnings_tuple[0]: + return states.RUNNING, None, 0 - if error_count > 0: - return states.ERROR, _failed_message() + if errors_tuples[0] > 0: + return states.ERROR, _failed_message(), 0 - return states.WAITING, _blocked_message() + # Remaining cardinality is just a difference between all tasks and + # a number of those tasks that induce RUNNING state. + cardinality = total_count - runnings_tuple[1] + + return states.WAITING, _blocked_message(), cardinality raise RuntimeError('Unexpected join expression: %s' % join_expr) @@ -337,51 +367,54 @@ class DirectWorkflowController(base.WorkflowController): def _get_induced_join_state(self, inbound_task_spec, join_task_spec): join_task_name = join_task_spec.get_name() - in_task_ex = self._find_task_execution_by_spec(inbound_task_spec) + in_task_ex = self._find_task_execution_by_name( + inbound_task_spec.get_name() + ) if not in_task_ex: - if self._possible_route(inbound_task_spec): - return states.WAITING + possible, depth = self._possible_route(inbound_task_spec) + + if possible: + return states.WAITING, depth else: - return states.ERROR + return states.ERROR, depth if not states.is_completed(in_task_ex.state): - return states.WAITING + return states.WAITING, 1 if join_task_name not in self._find_next_task_names(in_task_ex): - return states.ERROR + return states.ERROR, 1 - return states.RUNNING + return states.RUNNING, 1 - def _find_task_execution_by_spec(self, task_spec): - in_t_execs = wf_utils.find_task_executions_by_spec( - self.wf_ex, - task_spec + def _find_task_execution_by_name(self, t_name): + # Note: in case of 'join' completion check it's better to initialize + # the entire task_executions collection to avoid too many DB queries. + t_execs = lookup_utils.find_task_executions_by_name( + self.wf_ex.id, + t_name ) # TODO(rakhmerov): Temporary hack. See the previous comment. - return in_t_execs[-1] if in_t_execs else None + return t_execs[-1] if t_execs else None - def _possible_route(self, task_spec): - # TODO(rakhmerov): In some cases this method will be expensive because - # it uses a multistep recursive search with DB queries. - # It will be optimized with Workflow Execution Graph moving forward. + def _possible_route(self, task_spec, depth=1): in_task_specs = self.wf_spec.find_inbound_task_specs(task_spec) if not in_task_specs: - return True + return True, depth for t_s in in_task_specs: - t_ex = self._find_task_execution_by_spec(t_s) + t_ex = self._find_task_execution_by_name(t_s.get_name()) if not t_ex: - if self._possible_route(t_s): - return True + if self._possible_route(t_s, depth + 1): + return True, depth else: t_name = task_spec.get_name() if (not states.is_completed(t_ex.state) or t_name in self._find_next_task_names(t_ex)): - return True + return True, depth - return False + return False, depth diff --git a/mistral/workflow/lookup_utils.py b/mistral/workflow/lookup_utils.py new file mode 100644 index 000000000..9371f5491 --- /dev/null +++ b/mistral/workflow/lookup_utils.py @@ -0,0 +1,109 @@ +# Copyright 2015 - Mirantis, Inc. +# Copyright 2015 - StackStorm, Inc. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +""" +The intention of the module is providing various DB related lookup functions +for more convenient usage withing the workflow engine. + +Some of the functions may provide caching capabilities. + +WARNING: Oftentimes, persistent objects returned by the methods in this +module won't be attached to the current DB SQLAlchemy session because +they are returned from the cache and therefore they need to be used +carefully without trying to do any lazy loading etc. +These objects are also not suitable for re-attaching them to a session +in order to update their persistent DB state. +Mostly, they are useful for doing any kind of fast lookups with in order +to make some decision based on their state. +""" + +import cachetools +import threading + +from mistral.db.v2 import api as db_api +from mistral.workflow import states + +_TASK_EXECUTIONS_CACHE_LOCK = threading.RLock() +_TASK_EXECUTIONS_CACHE = cachetools.LRUCache(maxsize=20000) + + +def find_task_executions_by_name(wf_ex_id, task_name): + """Finds task executions by workflow execution id and task name. + + :param wf_ex_id: Workflow execution id. + :param task_name: Task name. + :return: Task executions (possibly a cached value). + """ + cache_key = (wf_ex_id, task_name) + + with _TASK_EXECUTIONS_CACHE_LOCK: + t_execs = _TASK_EXECUTIONS_CACHE.get(cache_key) + + if t_execs: + return t_execs + + t_execs = db_api.get_task_executions( + workflow_execution_id=wf_ex_id, + name=task_name + ) + + # We can cache only finished tasks because they won't change. + all_finished = ( + t_execs and + all([states.is_completed(t_ex.state) for t_ex in t_execs]) + ) + + if all_finished: + with _TASK_EXECUTIONS_CACHE_LOCK: + _TASK_EXECUTIONS_CACHE[cache_key] = t_execs + + return t_execs + + +def find_task_executions_by_spec(wf_ex_id, task_spec): + return find_task_executions_by_name(wf_ex_id, task_spec.get_name()) + + +def find_task_executions_by_specs(wf_ex_id, task_specs): + res = [] + + for t_s in task_specs: + res = res + find_task_executions_by_spec(wf_ex_id, t_s) + + return res + + +def find_task_executions_with_state(wf_ex_id, state): + return db_api.get_task_executions( + workflow_execution_id=wf_ex_id, + state=state + ) + + +def find_successful_task_executions(wf_ex_id): + return find_task_executions_with_state(wf_ex_id, states.SUCCESS) + + +def find_error_task_executions(wf_ex_id): + return find_task_executions_with_state(wf_ex_id, states.ERROR) + + +def find_cancelled_task_executions(wf_ex_id): + return find_task_executions_with_state(wf_ex_id, states.CANCELLED) + + +def clean_caches(): + with _TASK_EXECUTIONS_CACHE_LOCK: + _TASK_EXECUTIONS_CACHE.clear() diff --git a/mistral/workflow/reverse_workflow.py b/mistral/workflow/reverse_workflow.py index ebeb076eb..d4f2df755 100644 --- a/mistral/workflow/reverse_workflow.py +++ b/mistral/workflow/reverse_workflow.py @@ -19,8 +19,8 @@ from mistral import exceptions as exc from mistral.workflow import base from mistral.workflow import commands from mistral.workflow import data_flow +from mistral.workflow import lookup_utils from mistral.workflow import states -from mistral.workflow import utils as wf_utils class ReverseWorkflowController(base.WorkflowController): @@ -92,13 +92,16 @@ class ReverseWorkflowController(base.WorkflowController): return list( filter( lambda t_e: t_e.state == states.SUCCESS, - wf_utils.find_task_executions_by_specs(self.wf_ex, t_specs) + lookup_utils.find_task_executions_by_specs( + self.wf_ex.id, + t_specs + ) ) ) def evaluate_workflow_final_context(self): - task_execs = wf_utils.find_task_executions_by_spec( - self.wf_ex, + task_execs = lookup_utils.find_task_executions_by_spec( + self.wf_ex.id, self._get_target_task_specification() ) @@ -110,13 +113,15 @@ class ReverseWorkflowController(base.WorkflowController): def get_logical_task_state(self, task_ex): # TODO(rakhmerov): Implement. - return task_ex.state, task_ex.state_info + return task_ex.state, task_ex.state_info, 0 def is_error_handled_for(self, task_ex): return task_ex.state != states.ERROR def all_errors_handled(self): - return len(wf_utils.find_error_task_executions(self.wf_ex)) == 0 + task_execs = lookup_utils.find_error_task_executions(self.wf_ex.id) + + return len(task_execs) == 0 def _find_task_specs_with_satisfied_dependencies(self): """Given a target task name finds tasks with no dependencies. @@ -139,7 +144,8 @@ class ReverseWorkflowController(base.WorkflowController): ] def _is_satisfied_task(self, task_spec): - if wf_utils.find_task_executions_by_spec(self.wf_ex, task_spec): + if lookup_utils.find_task_executions_by_spec( + self.wf_ex.id, task_spec): return False if not self.wf_spec.get_task_requires(task_spec): diff --git a/mistral/workflow/utils.py b/mistral/workflow/utils.py index ed8f2b6af..7e955a122 100644 --- a/mistral/workflow/utils.py +++ b/mistral/workflow/utils.py @@ -14,9 +14,7 @@ # See the License for the specific language governing permissions and # limitations under the License. -from mistral.db.v2 import api as db_api from mistral.utils import serializers -from mistral.workflow import states class Result(object): @@ -72,46 +70,3 @@ class ResultSerializer(serializers.Serializer): entity['error'], entity.get('cancel', False) ) - - -def find_task_executions_by_name(wf_ex, task_name): - return db_api.get_task_executions( - workflow_execution_id=wf_ex.id, - name=task_name - ) - - -def find_task_executions_by_spec(wf_ex, task_spec): - return find_task_executions_by_name(wf_ex, task_spec.get_name()) - - -def find_task_executions_by_specs(wf_ex, task_specs): - res = [] - - for t_s in task_specs: - res = res + find_task_executions_by_spec(wf_ex, t_s) - - return res - - -def find_task_executions_with_state(wf_ex, state): - return db_api.get_task_executions( - workflow_execution_id=wf_ex.id, - state=state - ) - - -def find_running_task_executions(wf_ex): - return find_task_executions_with_state(wf_ex, states.RUNNING) - - -def find_successful_task_executions(wf_ex): - return find_task_executions_with_state(wf_ex, states.SUCCESS) - - -def find_error_task_executions(wf_ex): - return find_task_executions_with_state(wf_ex, states.ERROR) - - -def find_cancelled_task_executions(wf_ex): - return find_task_executions_with_state(wf_ex, states.CANCELLED)