Fix main process's reporting subprocess lost, switch back to thread mode

clearml/backend_interface/metrics/reporter.py

@@ -94,6 +94,26 @@ class BackgroundReportService(BackgroundMonitor, AsyncManagerMixin):
     def add_event(self, ev):
         if not self._queue:
             return
+        # check that we did not loose the reporter sub-process
+        if self.is_subprocess_mode() and not self._fast_is_subprocess_alive():
+            # we lost the reporting subprocess, let's switch to thread mode
+            # gel all data, work on local queue:
+            self._write()
+            # replace queue:
+            self._queue = TrQueue()
+            self._queue_size = 0
+            self._event = TrEvent()
+            self._done_ev = TrEvent()
+            self._start_ev = TrEvent()
+            self._exit_event = TrEvent()
+            self._empty_state_event = TrEvent()
+            self._res_waiting = Semaphore()
+            # set thread mode
+            self._subprocess = False
+            # start background thread
+            self._thread = None
+            self._start()
+
         self._queue.put(ev)
         self._queue_size += 1
         if self._queue_size >= self._flush_threshold:
@@ -103,12 +123,16 @@ class BackgroundReportService(BackgroundMonitor, AsyncManagerMixin):
         while not self._exit_event.wait(0):
             self._event.wait(self._wait_timeout)
             self._event.clear()
+            # lock state
             self._res_waiting.acquire()
             self._write()
             # wait for all reports
             if self.get_num_results() > 0:
                 self.wait_for_results()
-            self._empty_state_event.set()
+            # set empty flag only if we are not waiting for exit signal
+            if not self._exit_event.wait(0):
+                self._empty_state_event.set()
+            # unlock state
             self._res_waiting.release()
         # make sure we flushed everything
         self._async_enable = False

clearml/backend_interface/task/log.py

@@ -83,6 +83,23 @@ class BackgroundLogService(BackgroundMonitor):
         self._flush = SafeEvent()
 
     def add_to_queue(self, record):
+        # check that we did not loose the reporter sub-process
+        if self.is_subprocess_mode() and not self._fast_is_subprocess_alive():
+            # we lost the reporting subprocess, let's switch to thread mode
+            # gel all data, work on local queue:
+            self.send_all_records()
+            # replace queue:
+            self._queue = TrQueue()
+            self._flush = TrEvent()
+            self._event = TrEvent()
+            self._done_ev = TrEvent()
+            self._start_ev = TrEvent()
+            # set thread mode
+            self._subprocess = False
+            # start background thread
+            self._thread = None
+            self._start()
+
         self._queue.put(record)
 
     def empty(self):

clearml/utilities/process/mp.py

@@ -4,7 +4,7 @@ import struct
 import sys
 from functools import partial
 from multiprocessing import Lock, Event as ProcessEvent
-from threading import Thread, Event as TrEvent
+from threading import Thread, Event as TrEvent, RLock as ThreadRLock
 from time import sleep, time
 from typing import List, Dict, Optional
 from multiprocessing import Process
@@ -33,6 +33,39 @@ except ImportError:
         return False
 
 
+class ForkSafeRLock(object):
+    def __init__(self):
+        self._lock = None
+        self._instance_pid = None
+
+    def acquire(self, *args, **kwargs):
+        self.create()
+        return self._lock.acquire(*args, **kwargs)
+
+    def release(self, *args, **kwargs):
+        if self._lock is None:
+            return None
+        return self._lock.release(*args, **kwargs)
+
+    def create(self):
+        # this part is not atomic, and there is not a lot we can do about it.
+        if self._instance_pid != os.getpid() or not self._lock:
+            self._lock = ThreadRLock()
+            self._instance_pid = os.getpid()
+
+    def __enter__(self):
+        """Return `self` upon entering the runtime context."""
+        self.acquire()
+        return self
+
+    def __exit__(self, exc_type, exc_value, traceback):
+        """Raise any exception triggered within the runtime context."""
+        # Do whatever cleanup.
+        self.release()
+        if any((exc_type, exc_value, traceback,)):
+            raise (exc_type, exc_value, traceback)
+
+
 class ThreadCalls(object):
     def __init__(self):
         self._queue = TrQueue()
@@ -529,6 +562,16 @@ class BackgroundMonitor(object):
         else:
             return isinstance(self._thread, Thread) and self._thread.is_alive()
 
+    @classmethod
+    def _fast_is_subprocess_alive(cls):
+        if not cls._main_process:
+            return False
+        # noinspection PyBroadException
+        try:
+            return psutil.pid_exists(cls._main_process)
+        except Exception:
+            return False
+
     @classmethod
     def is_subprocess_alive(cls, task=None):
         if not cls._main_process or (task and cls._main_process_task_id != task.id):