Tweak how strategies evaluate failed hosts via the iterator and bug fixes
* Added additional methods to the iterator code to assess host failures while also taking into account the block rescue/always states * Fixed bugs in the free strategy, where results were not always being processed after being collected * Added some prettier printing to the state output from iterator Fixes #13699
This commit is contained in:
parent
6f2f7a79b3
commit
210cf06d9a
3 changed files with 49 additions and 14 deletions
|
@ -57,14 +57,32 @@ class HostState:
|
|||
self.always_child_state = None
|
||||
|
||||
def __repr__(self):
|
||||
return "HOST STATE: block=%d, task=%d, rescue=%d, always=%d, role=%s, run_state=%d, fail_state=%d, pending_setup=%s, tasks child state? %s, rescue child state? %s, always child state? %s" % (
|
||||
def _run_state_to_string(n):
|
||||
states = ["ITERATING_SETUP", "ITERATING_TASKS", "ITERATING_RESCUE", "ITERATING_ALWAYS", "ITERATING_COMPLETE"]
|
||||
try:
|
||||
return states[n]
|
||||
except IndexError:
|
||||
return "UNKNOWN STATE"
|
||||
|
||||
def _failed_state_to_string(n):
|
||||
states = {1:"FAILED_SETUP", 2:"FAILED_TASKS", 4:"FAILED_RESCUE", 8:"FAILED_ALWAYS"}
|
||||
if n == 0:
|
||||
return "FAILED_NONE"
|
||||
else:
|
||||
ret = []
|
||||
for i in (1, 2, 4, 8):
|
||||
if n & i:
|
||||
ret.append(states[i])
|
||||
return "|".join(ret)
|
||||
|
||||
return "HOST STATE: block=%d, task=%d, rescue=%d, always=%d, role=%s, run_state=%s, fail_state=%s, pending_setup=%s, tasks child state? %s, rescue child state? %s, always child state? %s" % (
|
||||
self.cur_block,
|
||||
self.cur_regular_task,
|
||||
self.cur_rescue_task,
|
||||
self.cur_always_task,
|
||||
self.cur_role,
|
||||
self.run_state,
|
||||
self.fail_state,
|
||||
_run_state_to_string(self.run_state),
|
||||
_failed_state_to_string(self.fail_state),
|
||||
self.pending_setup,
|
||||
self.tasks_child_state,
|
||||
self.rescue_child_state,
|
||||
|
@ -347,6 +365,28 @@ class PlayIterator:
|
|||
def get_failed_hosts(self):
|
||||
return dict((host, True) for (host, state) in iteritems(self._host_states) if state.run_state == self.ITERATING_COMPLETE and state.fail_state != self.FAILED_NONE)
|
||||
|
||||
def _check_failed_state(self, state):
|
||||
if state is None:
|
||||
return False
|
||||
elif state.run_state == self.ITERATING_TASKS and self._check_failed_state(state.tasks_child_state):
|
||||
return True
|
||||
elif state.run_state == self.ITERATING_RESCUE and self._check_failed_state(state.rescue_child_state):
|
||||
return True
|
||||
elif state.run_state == self.ITERATING_ALWAYS and self._check_failed_state(state.always_child_state):
|
||||
return True
|
||||
elif state.run_state == self.ITERATING_COMPLETE and state.fail_state != self.FAILED_NONE:
|
||||
if state.run_state == self.ITERATING_RESCUE and state.fail_state&self.FAILED_RESCUE == 0:
|
||||
return False
|
||||
elif state.run_state == self.ITERATING_ALWAYS and state.fail_state&self.FAILED_ALWAYS == 0:
|
||||
return False
|
||||
else:
|
||||
return True
|
||||
return False
|
||||
|
||||
def is_failed(self, host):
|
||||
s = self.get_host_state(host)
|
||||
return self._check_failed_state(s)
|
||||
|
||||
def get_original_task(self, host, task):
|
||||
'''
|
||||
Finds the task in the task list which matches the UUID of the given task.
|
||||
|
|
|
@ -78,7 +78,7 @@ class StrategyModule(StrategyBase):
|
|||
(state, task) = iterator.get_next_task_for_host(host, peek=True)
|
||||
display.debug("free host state: %s" % state)
|
||||
display.debug("free host task: %s" % task)
|
||||
if host_name not in self._tqm._failed_hosts and host_name not in self._tqm._unreachable_hosts and task:
|
||||
if not iterator.is_failed(host) and host_name not in self._tqm._unreachable_hosts and task:
|
||||
|
||||
# set the flag so the outer loop knows we've still found
|
||||
# some work which needs to be done
|
||||
|
@ -135,7 +135,7 @@ class StrategyModule(StrategyBase):
|
|||
if last_host == starting_host:
|
||||
break
|
||||
|
||||
results = self._process_pending_results(iterator)
|
||||
results = self._wait_on_pending_results(iterator)
|
||||
host_results.extend(results)
|
||||
|
||||
try:
|
||||
|
@ -176,13 +176,7 @@ class StrategyModule(StrategyBase):
|
|||
display.debug("done adding collected blocks to iterator")
|
||||
|
||||
# pause briefly so we don't spin lock
|
||||
time.sleep(0.05)
|
||||
|
||||
try:
|
||||
results = self._wait_on_pending_results(iterator)
|
||||
host_results.extend(results)
|
||||
except Exception as e:
|
||||
pass
|
||||
time.sleep(0.001)
|
||||
|
||||
# run the base class run() method, which executes the cleanup function
|
||||
# and runs any outstanding handlers which have been triggered
|
||||
|
|
|
@ -54,6 +54,7 @@ class StrategyModule(StrategyBase):
|
|||
host_tasks = {}
|
||||
display.debug("building list of next tasks for hosts")
|
||||
for host in hosts:
|
||||
if not iterator.is_failed(host):
|
||||
host_tasks[host.name] = iterator.get_next_task_for_host(host, peek=True)
|
||||
display.debug("done building task lists")
|
||||
|
||||
|
@ -98,7 +99,7 @@ class StrategyModule(StrategyBase):
|
|||
rvals = []
|
||||
display.debug("starting to advance hosts")
|
||||
for host in hosts:
|
||||
host_state_task = host_tasks[host.name]
|
||||
host_state_task = host_tasks.get(host.name)
|
||||
if host_state_task is None:
|
||||
continue
|
||||
(s, t) = host_state_task
|
||||
|
|
Loading…
Reference in a new issue