Remove defaults for Manager init, forcing keywords to be specified, a…

…nd perhaps revealign a bug that start_method is unused (#2973)

This PR removes theoretically-always-unused argument defaults in the process worker pool manager, and forces arguments to be supplied as kwargs.

These defaults have two downsides:

i) users reading the code might believe the defaults will be used as end-Parsl-user defaults. However, generally there will be up to two other layers of defaults that will override the defaults removed by this PR: defaults specified in the argparser of process_worker_pool; and defaults specified in the constructor to the HighThroughputExecutor object.

These defaults, then, are misleading, as they are defaults only in the presence of a coding error such as ii) below.

ii) in the presence of a coding error adding a new parameter, where the new parameter is not wired all the way through, rather than raise an error about a missing parameter, the otherwise-unused default is used. This was a practical problem demonstrated by the threads/spawn/fork start method parameter introduced in #2433 and removed as broken in #2980.

Switching to mandatory kwargs can help with accidentally misarranged parameters in the construction of a Manager object, a similar but different coding error to ii) above.

parsl/executors/high_throughput/process_worker_pool.py

@@ -48,22 +48,22 @@ class Manager:
                 |                          |                IPC-Qeueues
 
     """
-    def __init__(self,
-                 addresses="127.0.0.1",
-                 address_probe_timeout=30,
-                 task_port="50097",
-                 result_port="50098",
-                 cores_per_worker=1,
-                 mem_per_worker=None,
-                 max_workers=float('inf'),
-                 prefetch_capacity=0,
-                 uid=None,
-                 block_id=None,
-                 heartbeat_threshold=120,
-                 heartbeat_period=30,
-                 poll_period=10,
-                 cpu_affinity=False,
-                 available_accelerators: Sequence[str] = ()):
+    def __init__(self, *,
+                 addresses,
+                 address_probe_timeout,
+                 task_port,
+                 result_port,
+                 cores_per_worker,
+                 mem_per_worker,
+                 max_workers,
+                 prefetch_capacity,
+                 uid,
+                 block_id,
+                 heartbeat_threshold,
+                 heartbeat_period,
+                 poll_period,
+                 cpu_affinity,
+                 available_accelerators: Sequence[str]):
         """
         Parameters
         ----------
@@ -72,7 +72,7 @@ def __init__(self,
 
         address_probe_timeout : int
              Timeout in seconds for the address probe to detect viable addresses
-             to the interchange. Default : 30s
+             to the interchange.
 
         uid : str
              string unique identifier
@@ -82,43 +82,41 @@ def __init__(self,
 
         cores_per_worker : float
              cores to be assigned to each worker. Oversubscription is possible
-             by setting cores_per_worker < 1.0. Default=1
+             by setting cores_per_worker < 1.0.
 
         mem_per_worker : float
              GB of memory required per worker. If this option is specified, the node manager
              will check the available memory at startup and limit the number of workers such that
              the there's sufficient memory for each worker. If set to None, memory on node is not
              considered in the determination of workers to be launched on node by the manager.
-             Default: None
 
         max_workers : int
              caps the maximum number of workers that can be launched.
-             default: infinity
 
         prefetch_capacity : int
              Number of tasks that could be prefetched over available worker capacity.
              When there are a few tasks (<100) or when tasks are long running, this option should
-             be set to 0 for better load balancing. Default is 0.
+             be set to 0 for better load balancing.
 
         heartbeat_threshold : int
              Seconds since the last message from the interchange after which the
-             interchange is assumed to be un-available, and the manager initiates shutdown. Default:120s
+             interchange is assumed to be un-available, and the manager initiates shutdown.
 
              Number of seconds since the last message from the interchange after which the worker
-             assumes that the interchange is lost and the manager shuts down. Default:120
+             assumes that the interchange is lost and the manager shuts down.
 
         heartbeat_period : int
              Number of seconds after which a heartbeat message is sent to the interchange, and workers
              are checked for liveness.
 
         poll_period : int
-             Timeout period used by the manager in milliseconds. Default: 10ms
+             Timeout period used by the manager in milliseconds.
 
         cpu_affinity : str
              Whether or how each worker should force its affinity to different CPUs
 
         available_accelerators: list of str
-            List of accelerators available to the workers. Default: Empty list
+            List of accelerators available to the workers.
 
         """