Skip to content

Commit 56c07d2

Browse files
stepanblyschakstepanblyschak
authored
[healthd] fix healthd shutdown race (#19504)
#### Why I did it To fix errors that happen when writing to the queue: ``` Jun 5 23:04:41.798613 r-leopard-56 NOTICE healthd: Caught SIGTERM - exiting... Jun 5 23:04:41.798985 r-leopard-56 NOTICE healthd: Caught SIGTERM - exiting... Jun 5 23:04:41.799535 r-leopard-56 NOTICE healthd: Caught SIGTERM - exiting... Jun 5 23:04:41.806010 r-leopard-56 NOTICE healthd: Caught SIGTERM - exiting... Jun 5 23:04:41.814075 r-leopard-56 ERR healthd: system_service[Errno 104] Connection reset by peer Jun 5 23:04:41.824135 r-leopard-56 ERR healthd: Traceback (most recent call last):#12 File "/usr/local/lib/python3.9/dist-packages/health_checker/sysmonitor.py", line 484, in system_service#012 msg = self.myQ.get(timeout=QUEUE_TIMEOUT)#12 File "<string>", line 2, in get#012 File "/usr/lib/python3.9/multiprocessing/managers.py", line 809, in _callmethod#012 kind, result = conn.recv()#12 File "/usr/lib/python3.9/multiprocessing/connection.py", line 255, in recv#012 buf = self._recv_bytes()#12 File "/usr/lib/python3.9/multiprocessing/connection.py", line 419, in _recv_bytes#012 buf = self._recv(4)#12 File "/usr/lib/python3.9/multiprocessing/connection.py", line 384, in _recv#012 chunk = read(handle, remaining)#012ConnectionResetError: [Errno 104] Connection reset by peer Jun 5 23:04:41.826489 r-leopard-56 INFO healthd[8494]: ERROR:dbus.connection:Exception in handler for D-Bus signal: Jun 5 23:04:41.826591 r-leopard-56 INFO healthd[8494]: Traceback (most recent call last): Jun 5 23:04:41.826640 r-leopard-56 INFO healthd[8494]: File "/usr/lib/python3/dist-packages/dbus/connection.py", line 232, in maybe_handle_message Jun 5 23:04:41.826686 r-leopard-56 INFO healthd[8494]: self._handler(*args, **kwargs) Jun 5 23:04:41.826738 r-leopard-56 INFO healthd[8494]: File "/usr/local/lib/python3.9/dist-packages/health_checker/sysmonitor.py", line 82, in on_job_removed Jun 5 23:04:41.826785 r-leopard-56 INFO healthd[8494]: self.task_notify(msg) Jun 5 23:04:41.826831 r-leopard-56 INFO healthd[8494]: File "/usr/local/lib/python3.9/dist-packages/health_checker/sysmonitor.py", line 110, in task_notify Jun 5 23:04:41.826877 r-leopard-56 INFO healthd[8494]: self.task_queue.put(msg) Jun 5 23:04:41.826923 r-leopard-56 INFO healthd[8494]: File "<string>", line 2, in put Jun 5 23:04:41.826973 r-leopard-56 INFO healthd[8494]: File "/usr/lib/python3.9/multiprocessing/managers.py", line 808, in _callmethod Jun 5 23:04:41.827018 r-leopard-56 INFO healthd[8494]: conn.send((self._id, methodname, args, kwds)) Jun 5 23:04:41.827065 r-leopard-56 INFO healthd[8494]: File "/usr/lib/python3.9/multiprocessing/connection.py", line 211, in send Jun 5 23:04:41.827115 r-leopard-56 INFO healthd[8494]: self._send_bytes(_ForkingPickler.dumps(obj)) Jun 5 23:04:41.827158 r-leopard-56 INFO healthd[8494]: File "/usr/lib/python3.9/multiprocessing/connection.py", line 416, in _send_bytes Jun 5 23:04:41.827199 r-leopard-56 INFO healthd[8494]: self._send(header + buf) Jun 5 23:04:41.827254 r-leopard-56 INFO healthd[8494]: File "/usr/lib/python3.9/multiprocessing/connection.py", line 373, in _send Jun 5 23:04:41.827322 r-leopard-56 INFO healthd[8494]: n = write(self._handle, buf) Jun 5 23:04:41.827368 r-leopard-56 INFO healthd[8494]: BrokenPipeError: [Errno 32] Broken pipe Jun 5 23:04:42.800216 r-leopard-56 NOTICE healthd: Caught SIGTERM - exiting... ``` When the multiprocessing.Manager is shutdown the queue will raise the above errors. This happens during shutdown - fast-reboot, warm-reboot. With the fix, system-health service does not hang: ``` root@sonic:/home/admin# sudo systemctl start system-health ; sleep 10; echo "$(date): Stopping..."; sudo systemctl stop system-health; echo "$(date): Stopped" Thu Oct 17 01:07:56 PM IDT 2024: Stopping... Thu Oct 17 01:07:58 PM IDT 2024: Stopped root@sonic:/home/admin# sudo systemctl start system-health ; sleep 10; echo "$(date): Stopping..."; sudo systemctl stop system-health; echo "$(date): Stopped" Thu Oct 17 01:08:13 PM IDT 2024: Stopping... Thu Oct 17 01:08:14 PM IDT 2024: Stopped root@sonic:/home/admin# sudo systemctl start system-health ; sleep 10; echo "$(date): Stopping..."; sudo systemctl stop system-health; echo "$(date): Stopped" Thu Oct 17 01:09:05 PM IDT 2024: Stopping... Thu Oct 17 01:09:06 PM IDT 2024: Stopped ``` ##### Work item tracking - Microsoft ADO **(number only)**: #### How I did it Remove the call to shutdown, the cleanup will happen automatically when GC runs as per documentation - https://docs.python.org/3/library/multiprocessing.html #### How to verify it <!-- If PR needs to be backported, then the PR must be tested against the base branch and the earliest backport release branch and provide tested image version on these two branches. For example, if the PR is requested for master, 202211 and 202012, then the requester needs to provide test results on master and 202012. --> Run warm-reboot, fast-reboot multiple times and verify no errors in the log. #### Which release branch to backport (provide reason below if selected) <!-- - Note we only backport fixes to a release branch, *not* features! - Please also provide a reason for the backporting below. - e.g. - [x] 202006 --> - [ ] 201811 - [ ] 201911 - [ ] 202006 - [ ] 202012 - [ ] 202106 - [ ] 202111 - [x] 202205 - [x] 202311 - [x] 202405 #### Tested branch (Please provide the tested image version) <!-- - Please provide tested image version - e.g. - [x] 20201231.100 --> - [ ] <!-- image version 1 --> - [ ] <!-- image version 2 --> #### Description for the changelog <!-- Write a short (one line) summary that describes the changes in this pull request for inclusion in the changelog: --> <!-- Ensure to add label/tag for the feature raised. example - PR#2174 under sonic-utilities repo. where, Generic Config and Update feature has been labelled as GCU. --> #### Link to config_db schema for YANG module changes <!-- Provide a link to config_db schema for the table for which YANG model is defined Link should point to correct section on https://github.com/Azure/sonic-buildimage/blob/master/src/sonic-yang-models/doc/Configuration.md --> #### A picture of a cute animal (not mandatory but encouraged)
1 parent a98bee6 commit 56c07d2

File tree

1 file changed

+11
-13
lines changed

1 file changed

+11
-13
lines changed

‎src/system-health/health_checker/sysmonitor.py

+11-13
Original file line numberDiff line numberDiff line change
@@ -104,6 +104,11 @@ def task_worker(self):
104104
logger.log_info("Start Listening to systemd bus (pid {0})".format(os.getpid()))
105105
self.subscribe_sysbus()
106106

107+
def task_stop(self):
108+
# FIXME: Gracefully stop `loop.run()`.
109+
self._task_process.kill()
110+
return True
111+
107112
def task_notify(self, msg):
108113
if self.task_stopping_event.is_set():
109114
return
@@ -481,9 +486,11 @@ def system_service(self):
481486

482487
from queue import Empty
483488
# Queue to receive the STATEDB and Systemd state change event
484-
while not self.task_stopping_event.is_set():
489+
while True:
485490
try:
486491
msg = self.myQ.get(timeout=QUEUE_TIMEOUT)
492+
if msg == "stop":
493+
break
487494
event = msg["unit"]
488495
event_src = msg["evt_src"]
489496
event_time = msg["time"]
@@ -503,28 +510,19 @@ def system_service(self):
503510
monitor_statedb_table.task_stop()
504511

505512
def task_worker(self):
506-
if self.task_stopping_event.is_set():
507-
return
508513
self.system_service()
509514

510515
def task_stop(self):
511-
# Signal the process to stop
512-
self.task_stopping_event.set()
513-
#Clear the resources of mpmgr- Queue
514-
self.mpmgr.shutdown()
516+
self.myQ.put("stop")
515517

516518
# Wait for the process to exit
517519
self._task_process.join(self._stop_timeout_secs)
518520

519521
# If the process didn't exit, attempt to kill it
520522
if self._task_process.is_alive():
521523
logger.log_notice("Attempting to kill sysmon main process with pid {}".format(self._task_process.pid))
522-
os.kill(self._task_process.pid, signal.SIGKILL)
523-
524-
if self._task_process.is_alive():
525-
logger.log_error("Sysmon main process with pid {} could not be killed".format(self._task_process.pid))
524+
self._task_process.kill()
525+
self._task_process.join()
526526
return False
527527

528528
return True
529-
530-

0 commit comments

Comments
 (0)