Merge pull request #362 from pikers/ahab_you_bad_boi

Revert to hard container kill on log error
historical_breakeven_pp_price
goodboy 2022-07-21 17:51:11 -04:00 committed by GitHub
commit 0a7fce087c
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 16 additions and 3 deletions

View File

@ -204,9 +204,12 @@ class Container:
async def cancel( async def cancel(
self, self,
stop_msg: str, stop_msg: str,
hard_kill: bool = False,
) -> None: ) -> None:
cid = self.cntr.id cid = self.cntr.id
# first try a graceful cancel # first try a graceful cancel
log.cancel( log.cancel(
f'SIGINT cancelling container: {cid}\n' f'SIGINT cancelling container: {cid}\n'
@ -219,7 +222,12 @@ class Container:
with trio.move_on_after(0.5) as cs: with trio.move_on_after(0.5) as cs:
cs.shield = True cs.shield = True
await self.process_logs_until(stop_msg) log.cancel('polling for CNTR logs...')
try:
await self.process_logs_until(stop_msg)
except ApplicationLogError:
hard_kill = True
# if we aren't cancelled on above checkpoint then we # if we aren't cancelled on above checkpoint then we
# assume we read the expected stop msg and terminated. # assume we read the expected stop msg and terminated.
@ -234,6 +242,7 @@ class Container:
condition='not-running', condition='not-running',
) )
# graceful exit if we didn't time out
break break
except ( except (
@ -249,9 +258,12 @@ class Container:
log.exception('Docker connection failure') log.exception('Docker connection failure')
self.hard_kill(start) self.hard_kill(start)
else: else:
self.hard_kill(start) hard_kill = True
log.cancel(f'Container stopped: {cid}') if hard_kill:
self.hard_kill(start)
else:
log.cancel(f'Container stopped: {cid}')
@tractor.context @tractor.context
@ -300,6 +312,7 @@ async def open_ahabd(
await trio.sleep_forever() await trio.sleep_forever()
finally: finally:
# needed?
with trio.CancelScope(shield=True): with trio.CancelScope(shield=True):
await cntr.cancel(stop_msg) await cntr.cancel(stop_msg)