parsl 2024.10.14__py3-none-any.whl → 2024.10.21__py3-none-any.whl

parsl/executors/high_throughput/executor.py

@@ -146,6 +146,11 @@ GENERAL_HTEX_PARAM_DOCS = """provider : :class:`~parsl.providers.base.ExecutionP
 
     encrypted : bool
         Flag to enable/disable encryption (CurveZMQ). Default is False.
+
+    manager_selector: ManagerSelector
+        Determines what strategy the interchange uses to select managers during task distribution.
+        See API reference under "Manager Selectors" regarding the various manager selectors.
+        Default: 'RandomManagerSelector'
 """  # Documentation for params used by both HTEx and MPIEx
 
 
@@ -341,15 +346,17 @@ class HighThroughputExecutor(BlockProviderExecutor, RepresentationMixin, UsageIn
         return self.logdir
 
     def validate_resource_spec(self, resource_specification: dict):
-        """HTEX does not support *any* resource_specification options and
-        will raise InvalidResourceSpecification is any are passed to it"""
+        """HTEX supports the following *Optional* resource specifications:
+        priority: lower value is higher priority"""
         if resource_specification:
-            raise InvalidResourceSpecification(
-                set(resource_specification.keys()),
-                ("HTEX does not support the supplied resource_specifications."
-                 "For MPI applications consider using the MPIExecutor. "
-                 "For specifications for core count/memory/walltime, consider using WorkQueueExecutor. ")
-            )
+            acceptable_fields = {'priority'}
+            keys = set(resource_specification.keys())
+            invalid_keys = keys - acceptable_fields
+            if invalid_keys:
+                message = "Task resource specification only accepts these types of resources: {}".format(
+                    ', '.join(acceptable_fields))
+                logger.error(message)
+                raise InvalidResourceSpecification(set(invalid_keys), message)
         return
 
     def initialize_scaling(self):
@@ -657,7 +664,7 @@ class HighThroughputExecutor(BlockProviderExecutor, RepresentationMixin, UsageIn
         except TypeError:
             raise SerializationError(func.__name__)
 
-        msg = {"task_id": task_id, "buffer": fn_buf}
+        msg = {"task_id": task_id, "resource_spec": resource_specification, "buffer": fn_buf}
 
         # Post task to the outgoing queue
         self.outgoing_q.put(msg)

parsl/executors/high_throughput/manager_selector.py

@@ -19,7 +19,37 @@ class ManagerSelector(metaclass=ABCMeta):
 
 class RandomManagerSelector(ManagerSelector):
 
+    """Returns a shuffled list of interesting_managers
+
+    By default this strategy is used by the interchange. Works well
+    in distributing workloads equally across all availble compute
+    resources. The random workload strategy is not effective in
+    conjunction with elastic scaling behavior as the even task
+    distribution does not allow the scaling down of blocks, leading
+    to wasted resource consumption.
+    """
+
     def sort_managers(self, ready_managers: Dict[bytes, ManagerRecord], manager_list: Set[bytes]) -> List[bytes]:
         c_manager_list = list(manager_list)
         random.shuffle(c_manager_list)
         return c_manager_list
+
+
+class BlockIdManagerSelector(ManagerSelector):
+
+    """Returns an interesting_managers list sorted by block ID
+
+    Observations:
+    1. BlockID manager selector helps with workloads that see a varying
+    amount of tasks over time. New blocks are prioritized with the
+    blockID manager selector, when used with 'htex_auto_scaling', results
+    in compute cost savings.
+
+    2. Doesn't really work with bag-of-tasks workloads. When all the tasks
+    are put into the queue upfront, all blocks operate at near full
+    utilization for the majority of the workload, which task goes where
+    doesn't really matter.
+    """
+
+    def sort_managers(self, ready_managers: Dict[bytes, ManagerRecord], manager_list: Set[bytes]) -> List[bytes]:
+        return sorted(manager_list, key=lambda x: (ready_managers[x]['block_id'] is not None, ready_managers[x]['block_id']))

parsl/executors/high_throughput/process_worker_pool.py

@@ -362,7 +362,7 @@ class Manager:
                 if tasks == HEARTBEAT_CODE:
                     logger.debug("Got heartbeat from interchange")
                 elif tasks == DRAINED_CODE:
-                    logger.info("Got fulled drained message from interchange - setting kill flag")
+                    logger.info("Got fully drained message from interchange - setting kill flag")
                     kill_event.set()
                 else:
                     task_recv_counter += len(tasks)

parsl/providers/kubernetes/kube.py

@@ -1,10 +1,5 @@
 import logging
-import time
-
-from parsl.providers.kubernetes.template import template_string
-
-logger = logging.getLogger(__name__)
-
+import uuid
 from typing import Any, Dict, List, Optional, Tuple
 
 import typeguard
@@ -12,7 +7,8 @@ import typeguard
 from parsl.errors import OptionalModuleMissing
 from parsl.jobs.states import JobState, JobStatus
 from parsl.providers.base import ExecutionProvider
-from parsl.utils import RepresentationMixin
+from parsl.providers.kubernetes.template import template_string
+from parsl.utils import RepresentationMixin, sanitize_dns_subdomain_rfc1123
 
 try:
     from kubernetes import client, config
@@ -20,6 +16,8 @@ try:
 except (ImportError, NameError, FileNotFoundError):
     _kubernetes_enabled = False
 
+logger = logging.getLogger(__name__)
+
 translate_table = {
     'Running': JobState.RUNNING,
     'Pending': JobState.PENDING,
@@ -161,7 +159,7 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
         self.resources: Dict[object, Dict[str, Any]]
         self.resources = {}
 
-    def submit(self, cmd_string, tasks_per_node, job_name="parsl"):
+    def submit(self, cmd_string: str, tasks_per_node: int, job_name: str = "parsl.kube"):
         """ Submit a job
         Args:
              - cmd_string  :(String) - Name of the container to initiate
@@ -173,15 +171,19 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
         Returns:
              - job_id: (string) Identifier for the job
         """
+        job_id = uuid.uuid4().hex[:8]
 
-        cur_timestamp = str(time.time() * 1000).split(".")[0]
-        job_name = "{0}-{1}".format(job_name, cur_timestamp)
-
-        if not self.pod_name:
-            pod_name = '{}'.format(job_name)
-        else:
-            pod_name = '{}-{}'.format(self.pod_name,
-                                      cur_timestamp)
+        pod_name = self.pod_name or job_name
+        try:
+            pod_name = sanitize_dns_subdomain_rfc1123(pod_name)
+        except ValueError:
+            logger.warning(
+                f"Invalid pod name '{pod_name}' for job '{job_id}', falling back to 'parsl.kube'"
+            )
+            pod_name = "parsl.kube"
+        pod_name = pod_name[:253 - 1 - len(job_id)]  # Leave room for the job ID
+        pod_name = pod_name.rstrip(".-")  # Remove trailing dot or hyphen after trim
+        pod_name = f"{pod_name}.{job_id}"
 
         formatted_cmd = template_string.format(command=cmd_string,
                                                worker_init=self.worker_init)
@@ -189,14 +191,14 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
         logger.debug("Pod name: %s", pod_name)
         self._create_pod(image=self.image,
                          pod_name=pod_name,
-                         job_name=job_name,
+                         job_id=job_id,
                          cmd_string=formatted_cmd,
                          volumes=self.persistent_volumes,
                          service_account_name=self.service_account_name,
                          annotations=self.annotations)
-        self.resources[pod_name] = {'status': JobStatus(JobState.RUNNING)}
+        self.resources[job_id] = {'status': JobStatus(JobState.RUNNING), 'pod_name': pod_name}
 
-        return pod_name
+        return job_id
 
     def status(self, job_ids):
         """ Get the status of a list of jobs identified by the job identifiers
@@ -212,6 +214,9 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
             self._status()
         return [self.resources[jid]['status'] for jid in job_ids]
 
+    def _get_pod_name(self, job_id: str) -> str:
+        return self.resources[job_id]['pod_name']
+
     def cancel(self, job_ids):
         """ Cancels the jobs specified by a list of job ids
         Args:
@@ -221,7 +226,8 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
         """
         for job in job_ids:
             logger.debug("Terminating job/pod: {0}".format(job))
-            self._delete_pod(job)
+            pod_name = self._get_pod_name(job)
+            self._delete_pod(pod_name)
 
             self.resources[job]['status'] = JobStatus(JobState.CANCELLED)
         rets = [True for i in job_ids]
@@ -242,7 +248,8 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
         for jid in to_poll_job_ids:
             phase = None
             try:
-                pod = self.kube_client.read_namespaced_pod(name=jid, namespace=self.namespace)
+                pod_name = self._get_pod_name(jid)
+                pod = self.kube_client.read_namespaced_pod(name=pod_name, namespace=self.namespace)
             except Exception:
                 logger.exception("Failed to poll pod {} status, most likely because pod was terminated".format(jid))
                 if self.resources[jid]['status'] is JobStatus(JobState.RUNNING):
@@ -257,10 +264,10 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
                 self.resources[jid]['status'] = JobStatus(status)
 
     def _create_pod(self,
-                    image,
-                    pod_name,
-                    job_name,
-                    port=80,
+                    image: str,
+                    pod_name: str,
+                    job_id: str,
+                    port: int = 80,
                     cmd_string=None,
                     volumes=[],
                     service_account_name=None,
@@ -269,7 +276,7 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
         Args:
               - image (string) : Docker image to launch
               - pod_name (string) : Name of the pod
-              - job_name (string) : App label
+              - job_id (string) : Job ID
         KWargs:
              - port (integer) : Container port
         Returns:
@@ -299,7 +306,7 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
                                                   )
         # Configure Pod template container
         container = client.V1Container(
-            name=pod_name,
+            name=job_id,
             image=image,
             resources=resources,
             ports=[client.V1ContainerPort(container_port=port)],
@@ -322,7 +329,7 @@ class KubernetesProvider(ExecutionProvider, RepresentationMixin):
                                                    claim_name=volume[0])))
 
         metadata = client.V1ObjectMeta(name=pod_name,
-                                       labels={"app": job_name},
+                                       labels={"parsl-job-id": job_id},
                                        annotations=annotations)
         spec = client.V1PodSpec(containers=[container],
                                 image_pull_secrets=[secret],

parsl/tests/test_htex/test_block_manager_selector_unit.py

@@ -0,0 +1,20 @@
+import pytest
+
+from parsl.executors.high_throughput.manager_record import ManagerRecord
+from parsl.executors.high_throughput.manager_selector import BlockIdManagerSelector
+
+
+@pytest.mark.local
+def test_sort_managers():
+    ready_managers = {
+        b'manager1': {'block_id': 1},
+        b'manager2': {'block_id': None},
+        b'manager3': {'block_id': 3},
+        b'manager4': {'block_id': 2}
+    }
+
+    manager_list = {b'manager1', b'manager2', b'manager3', b'manager4'}
+    expected_sorted_list = [b'manager2', b'manager1', b'manager4', b'manager3']
+    manager_selector = BlockIdManagerSelector()
+    sorted_managers = manager_selector.sort_managers(ready_managers, manager_list)
+    assert sorted_managers == expected_sorted_list

parsl/tests/test_htex/test_drain.py

@@ -13,7 +13,9 @@ from parsl.providers import LocalProvider
 # based around the expected drain period: the drain period
 # is TIME_CONST seconds, and the single executed task will
 # last twice that many number of seconds.
-TIME_CONST = 1
+TIME_CONST = 4
+
+CONNECTED_MANAGERS_POLL_MS = 100
 
 
 def local_config():
@@ -52,7 +54,7 @@ def test_drain(try_assert):
 
     # wait till we have a block running...
 
-    try_assert(lambda: len(htex.connected_managers()) == 1)
+    try_assert(lambda: len(htex.connected_managers()) == 1, check_period_ms=CONNECTED_MANAGERS_POLL_MS)
 
     managers = htex.connected_managers()
     assert managers[0]['active'], "The manager should be active"
@@ -63,7 +65,7 @@ def test_drain(try_assert):
     time.sleep(TIME_CONST)
 
     # this assert should happen *very fast* after the above delay...
-    try_assert(lambda: htex.connected_managers()[0]['draining'], timeout_ms=500)
+    try_assert(lambda: htex.connected_managers()[0]['draining'], timeout_ms=500, check_period_ms=CONNECTED_MANAGERS_POLL_MS)
 
     # and the test task should still be running...
     assert not fut.done(), "The test task should still be running"
@@ -76,4 +78,4 @@ def test_drain(try_assert):
     # connected managers.
     # As with the above draining assert, this should happen very fast after
     # the task ends.
-    try_assert(lambda: len(htex.connected_managers()) == 0, timeout_ms=500)
+    try_assert(lambda: len(htex.connected_managers()) == 0, timeout_ms=500, check_period_ms=CONNECTED_MANAGERS_POLL_MS)

parsl/tests/test_htex/test_manager_selector_by_block.py

@@ -0,0 +1,53 @@
+import time
+
+import pytest
+
+import parsl
+from parsl.app.app import bash_app, python_app
+from parsl.channels import LocalChannel
+from parsl.config import Config
+from parsl.executors import HighThroughputExecutor
+from parsl.executors.high_throughput.manager_selector import (
+    BlockIdManagerSelector,
+    ManagerSelector,
+)
+from parsl.launchers import WrappedLauncher
+from parsl.providers import LocalProvider
+from parsl.usage_tracking.levels import LEVEL_1
+
+BLOCK_COUNT = 2
+
+
+@parsl.python_app
+def get_worker_pid():
+    import os
+    return os.environ.get('PARSL_WORKER_BLOCK_ID')
+
+
+@pytest.mark.local
+def test_block_id_selection(try_assert):
+    htex = HighThroughputExecutor(
+        label="htex_local",
+        max_workers_per_node=1,
+        manager_selector=BlockIdManagerSelector(),
+        provider=LocalProvider(
+            channel=LocalChannel(),
+            init_blocks=BLOCK_COUNT,
+            max_blocks=BLOCK_COUNT,
+            min_blocks=BLOCK_COUNT,
+        ),
+    )
+
+    config = Config(
+        executors=[htex],
+        usage_tracking=LEVEL_1,
+    )
+
+    with parsl.load(config):
+        blockids = []
+        try_assert(lambda: len(htex.connected_managers()) == BLOCK_COUNT, timeout_ms=20000)
+        for i in range(10):
+            future = get_worker_pid()
+            blockids.append(future.result())
+
+        assert all(blockid == "1" for blockid in blockids)

parsl/tests/test_htex/test_resource_spec_validation.py

@@ -30,6 +30,13 @@ def test_resource_spec_validation():
     assert ret_val is None
 
 
+@pytest.mark.local
+def test_resource_spec_validation_one_key():
+    htex = HighThroughputExecutor()
+    ret_val = htex.validate_resource_spec({"priority": 2})
+    assert ret_val is None
+
+
 @pytest.mark.local
 def test_resource_spec_validation_bad_keys():
     htex = HighThroughputExecutor()

parsl/tests/test_providers/test_kubernetes_provider.py

@@ -0,0 +1,102 @@
+import re
+from unittest import mock
+
+import pytest
+
+from parsl.providers.kubernetes.kube import KubernetesProvider
+from parsl.tests.test_utils.test_sanitize_dns import DNS_SUBDOMAIN_REGEX
+
+_MOCK_BASE = "parsl.providers.kubernetes.kube"
+
+
+@pytest.fixture(autouse=True)
+def mock_kube_config():
+    with mock.patch(f"{_MOCK_BASE}.config") as mock_config:
+        mock_config.load_kube_config.return_value = None
+        yield mock_config
+
+
+@pytest.fixture
+def mock_kube_client():
+    mock_client = mock.MagicMock()
+    with mock.patch(f"{_MOCK_BASE}.client.CoreV1Api") as mock_api:
+        mock_api.return_value = mock_client
+        yield mock_client
+
+
+@pytest.mark.local
+def test_submit_happy_path(mock_kube_client: mock.MagicMock):
+    image = "test-image"
+    namespace = "test-namespace"
+    cmd_string = "test-command"
+    volumes = [("test-volume", "test-mount-path")]
+    service_account_name = "test-service-account"
+    annotations = {"test-annotation": "test-value"}
+    max_cpu = 2
+    max_mem = "2Gi"
+    init_cpu = 1
+    init_mem = "1Gi"
+    provider = KubernetesProvider(
+        image=image,
+        persistent_volumes=volumes,
+        namespace=namespace,
+        service_account_name=service_account_name,
+        annotations=annotations,
+        max_cpu=max_cpu,
+        max_mem=max_mem,
+        init_cpu=init_cpu,
+        init_mem=init_mem,
+    )
+
+    job_name = "test.job.name"
+    job_id = provider.submit(cmd_string=cmd_string, tasks_per_node=1, job_name=job_name)
+
+    assert job_id in provider.resources
+    assert mock_kube_client.create_namespaced_pod.call_count == 1
+
+    call_args = mock_kube_client.create_namespaced_pod.call_args[1]
+    pod = call_args["body"]
+    container = pod.spec.containers[0]
+    volume = container.volume_mounts[0]
+
+    assert image == container.image
+    assert namespace == call_args["namespace"]
+    assert any(cmd_string in arg for arg in container.args)
+    assert volumes[0] == (volume.name, volume.mount_path)
+    assert service_account_name == pod.spec.service_account_name
+    assert annotations == pod.metadata.annotations
+    assert str(max_cpu) == container.resources.limits["cpu"]
+    assert max_mem == container.resources.limits["memory"]
+    assert str(init_cpu) == container.resources.requests["cpu"]
+    assert init_mem == container.resources.requests["memory"]
+    assert job_id == pod.metadata.labels["parsl-job-id"]
+    assert job_id == container.name
+    assert f"{job_name}.{job_id}" == pod.metadata.name
+
+
+@pytest.mark.local
+@mock.patch(f"{_MOCK_BASE}.KubernetesProvider._create_pod")
+@pytest.mark.parametrize("char", (".", "-"))
+def test_submit_pod_name_includes_job_id(mock_create_pod: mock.MagicMock, char: str):
+    provider = KubernetesProvider(image="test-image")
+
+    job_name = "a." * 121 + f"a{char}" + "a" * 9
+    assert len(job_name) == 253  # Max length for pod name
+    job_id = provider.submit(cmd_string="test-command", tasks_per_node=1, job_name=job_name)
+
+    expected_pod_name = job_name[:253 - len(job_id) - 2] + f".{job_id}"
+    actual_pod_name = mock_create_pod.call_args[1]["pod_name"]
+    assert re.match(DNS_SUBDOMAIN_REGEX, actual_pod_name)
+    assert expected_pod_name == actual_pod_name
+
+
+@pytest.mark.local
+@mock.patch(f"{_MOCK_BASE}.KubernetesProvider._create_pod")
+@mock.patch(f"{_MOCK_BASE}.logger")
+@pytest.mark.parametrize("job_name", ("", ".", "-", "a.-.a", "$$$"))
+def test_submit_invalid_job_name(mock_logger: mock.MagicMock, mock_create_pod: mock.MagicMock, job_name: str):
+    provider = KubernetesProvider(image="test-image")
+    job_id = provider.submit(cmd_string="test-command", tasks_per_node=1, job_name=job_name)
+    assert mock_logger.warning.call_count == 1
+    assert f"Invalid pod name '{job_name}' for job '{job_id}'" in mock_logger.warning.call_args[0][0]
+    assert f"parsl.kube.{job_id}" == mock_create_pod.call_args[1]["pod_name"]

parsl/tests/test_utils/test_sanitize_dns.py

@@ -0,0 +1,76 @@
+import random
+import re
+
+import pytest
+
+from parsl.utils import sanitize_dns_label_rfc1123, sanitize_dns_subdomain_rfc1123
+
+# Ref: https://datatracker.ietf.org/doc/html/rfc1123
+DNS_LABEL_REGEX = r'^[a-z0-9]([-a-z0-9]{0,61}[a-z0-9])?$'
+DNS_SUBDOMAIN_REGEX = r'^[a-z0-9]([-a-z0-9]{0,61}[a-z0-9])?(\.[a-z0-9]([-a-z0-9]{0,61}[a-z0-9])?)*$'
+
+test_labels = [
+    "example-label-123",           # Valid label
+    "EXAMPLE",                     # Case sensitivity
+    "!@#example*",                 # Remove invalid characters
+    "--leading-and-trailing--",    # Leading and trailing hyphens
+    "..leading.and.trailing..",    # Leading and tailing dots
+    "multiple..dots",              # Consecutive dots
+    "valid--label",                # Consecutive hyphens
+    "a" * random.randint(64, 70),  # Longer than 63 characters
+    f"{'a' * 62}-a",               # Trailing hyphen at max length
+]
+
+
+def _generate_test_subdomains(num_subdomains: int):
+    subdomains = []
+    for _ in range(num_subdomains):
+        num_labels = random.randint(1, 5)
+        labels = [test_labels[random.randint(0, num_labels - 1)] for _ in range(num_labels)]
+        subdomain = ".".join(labels)
+        subdomains.append(subdomain)
+    return subdomains
+
+
+@pytest.mark.local
+@pytest.mark.parametrize("raw_string", test_labels)
+def test_sanitize_dns_label_rfc1123(raw_string: str):
+    print(sanitize_dns_label_rfc1123(raw_string))
+    assert re.match(DNS_LABEL_REGEX, sanitize_dns_label_rfc1123(raw_string))
+
+
+@pytest.mark.local
+@pytest.mark.parametrize("raw_string", ("", "-", "@", "$$$"))
+def test_sanitize_dns_label_rfc1123_empty(raw_string: str):
+    with pytest.raises(ValueError) as e_info:
+        sanitize_dns_label_rfc1123(raw_string)
+    assert str(e_info.value) == f"Sanitized DNS label is empty for input '{raw_string}'"
+
+
+@pytest.mark.local
+@pytest.mark.parametrize("raw_string", _generate_test_subdomains(10))
+def test_sanitize_dns_subdomain_rfc1123(raw_string: str):
+    assert re.match(DNS_SUBDOMAIN_REGEX, sanitize_dns_subdomain_rfc1123(raw_string))
+
+
+@pytest.mark.local
+@pytest.mark.parametrize("char", ("-", "."))
+def test_sanitize_dns_subdomain_rfc1123_trailing_non_alphanumeric_at_max_length(char: str):
+    raw_string = (f"{'a' * 61}." * 4) + f".aaaa{char}a"
+    assert re.match(DNS_SUBDOMAIN_REGEX, sanitize_dns_subdomain_rfc1123(raw_string))
+
+
+@pytest.mark.local
+@pytest.mark.parametrize("raw_string", ("", ".", "..."))
+def test_sanitize_dns_subdomain_rfc1123_empty(raw_string: str):
+    with pytest.raises(ValueError) as e_info:
+        sanitize_dns_subdomain_rfc1123(raw_string)
+    assert str(e_info.value) == f"Sanitized DNS subdomain is empty for input '{raw_string}'"
+
+
+@pytest.mark.local
+@pytest.mark.parametrize(
+    "raw_string", ("a" * 253, "a" * random.randint(254, 300)), ids=("254 chars", ">253 chars")
+)
+def test_sanitize_dns_subdomain_rfc1123_max_length(raw_string: str):
+    assert len(sanitize_dns_subdomain_rfc1123(raw_string)) <= 253

parsl/utils.py

@@ -1,6 +1,7 @@
 import inspect
 import logging
 import os
+import re
 import shlex
 import subprocess
 import threading
@@ -380,3 +381,80 @@ class AutoCancelTimer(threading.Timer):
         exc_tb: Optional[TracebackType]
     ) -> None:
         self.cancel()
+
+
+def sanitize_dns_label_rfc1123(raw_string: str) -> str:
+    """Convert input string to a valid RFC 1123 DNS label.
+
+    Parameters
+    ----------
+    raw_string : str
+        String to sanitize.
+
+    Returns
+    -------
+    str
+        Sanitized string.
+
+    Raises
+    ------
+    ValueError
+        If the string is empty after sanitization.
+    """
+    # Convert to lowercase and replace non-alphanumeric characters with hyphen
+    sanitized = re.sub(r'[^a-z0-9]', '-', raw_string.lower())
+
+    # Remove consecutive hyphens
+    sanitized = re.sub(r'-+', '-', sanitized)
+
+    # DNS label cannot exceed 63 characters
+    sanitized = sanitized[:63]
+
+    # Strip after trimming to avoid trailing hyphens
+    sanitized = sanitized.strip("-")
+
+    if not sanitized:
+        raise ValueError(f"Sanitized DNS label is empty for input '{raw_string}'")
+
+    return sanitized
+
+
+def sanitize_dns_subdomain_rfc1123(raw_string: str) -> str:
+    """Convert input string to a valid RFC 1123 DNS subdomain.
+
+    Parameters
+    ----------
+    raw_string : str
+        String to sanitize.
+
+    Returns
+    -------
+    str
+        Sanitized string.
+
+    Raises
+    ------
+    ValueError
+        If the string is empty after sanitization.
+    """
+    segments = raw_string.split('.')
+
+    sanitized_segments = []
+    for segment in segments:
+        if not segment:
+            continue
+        sanitized_segment = sanitize_dns_label_rfc1123(segment)
+        sanitized_segments.append(sanitized_segment)
+
+    sanitized = '.'.join(sanitized_segments)
+
+    # DNS subdomain cannot exceed 253 characters
+    sanitized = sanitized[:253]
+
+    # Strip after trimming to avoid trailing dots or hyphens
+    sanitized = sanitized.strip(".-")
+
+    if not sanitized:
+        raise ValueError(f"Sanitized DNS subdomain is empty for input '{raw_string}'")
+
+    return sanitized

parsl/version.py

@@ -3,4 +3,4 @@
 Year.Month.Day[alpha/beta/..]
 Alphas will be numbered like this -> 2024.12.10a0
 """
-VERSION = '2024.10.14'
+VERSION = '2024.10.21'

{parsl-2024.10.14.data → parsl-2024.10.21.data}/scripts/process_worker_pool.py

@@ -362,7 +362,7 @@ class Manager:
                 if tasks == HEARTBEAT_CODE:
                     logger.debug("Got heartbeat from interchange")
                 elif tasks == DRAINED_CODE:
-                    logger.info("Got fulled drained message from interchange - setting kill flag")
+                    logger.info("Got fully drained message from interchange - setting kill flag")
                     kill_event.set()
                 else:
                     task_recv_counter += len(tasks)

{parsl-2024.10.14.dist-info → parsl-2024.10.21.dist-info}/METADATA

@@ -1,9 +1,9 @@
 Metadata-Version: 2.1
 Name: parsl
-Version: 2024.10.14
+Version: 2024.10.21
 Summary: Simple data dependent workflows in Python
 Home-page: https://github.com/Parsl/parsl
-Download-URL: https://github.com/Parsl/parsl/archive/2024.10.14.tar.gz
+Download-URL: https://github.com/Parsl/parsl/archive/2024.10.21.tar.gz
 Author: The Parsl Team
 Author-email: parsl@googlegroups.com
 License: Apache 2.0

{parsl-2024.10.14.dist-info → parsl-2024.10.21.dist-info}/RECORD

@@ -7,8 +7,8 @@ parsl/log_utils.py,sha256=u14Fkl5eDfS4HMpl0JjseNNPdbvPaugWPRQj1_af_Zo,3273
 parsl/multiprocessing.py,sha256=MyaEcEq-Qf860u7V98u-PZrPNdtzOZL_NW6EhIJnmfQ,1937
 parsl/process_loggers.py,sha256=uQ7Gd0W72Jz7rrcYlOMfLsAEhkRltxXJL2MgdduJjEw,1136
 parsl/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-parsl/utils.py,sha256=91FjQiTUY383ueAjkBAgE21My9nba6SP2a2SrbB1r1Q,11250
-parsl/version.py,sha256=gXe4u5wct19WXUeUe498E2p6fv6jhUpyBFJ0x-XoQdM,131
+parsl/utils.py,sha256=rMLKeadEsQ9jGwm4ogqiLIXPS3zOAyfznQJXVkJSY8E,13107
+parsl/version.py,sha256=0V6_ogkULPZVJXRQqKVT9TwsP2SpvX2cDNjSb1ouhPk,131
 parsl/app/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/app/app.py,sha256=0gbM4AH2OtFOLsv07I5nglpElcwMSOi-FzdZZfrk7So,8532
 parsl/app/bash.py,sha256=jm2AvePlCT9DZR7H_4ANDWxatp5dN_22FUlT_gWhZ-g,5528
@@ -80,16 +80,16 @@ parsl/executors/flux/executor.py,sha256=8_xakLUu5zNJAHL0LbeTCFEWqWzRK1eE-3ep4GII
 parsl/executors/flux/flux_instance_manager.py,sha256=5T3Rp7ZM-mlT0Pf0Gxgs5_YmnaPrSF9ec7zvRfLfYJw,2129
 parsl/executors/high_throughput/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/executors/high_throughput/errors.py,sha256=Sak8e8UpiEcXefUjMHbhyXc4Rn7kJtOoh7L8wreBQdk,1638
-parsl/executors/high_throughput/executor.py,sha256=5UGCQR3OtTXXMMo7dm0n80HNc0iqZ30NlU7sLlAdTf8,37139
+parsl/executors/high_throughput/executor.py,sha256=_dff5USFQq7V89kEXEWd2OqgYJQfq9i1b2e8FYA-zow,37511
 parsl/executors/high_throughput/interchange.py,sha256=elt48I-3WI4Wf5s7_3ECTw_fqqLPBDA2IzOiC4vqB14,29925
 parsl/executors/high_throughput/manager_record.py,sha256=yn3L8TUJFkgm2lX1x0SeS9mkvJowC0s2VIMCFiU7ThM,455
-parsl/executors/high_throughput/manager_selector.py,sha256=uRaEtcbDO2vXf8vjEcm7bfZVdeUlSPTRc3G4oFRO29M,820
+parsl/executors/high_throughput/manager_selector.py,sha256=UKcUE6v0tO7PDMTThpKSKxVpOpOUilxDL7UbNgpZCxo,2116
 parsl/executors/high_throughput/monitoring_info.py,sha256=HC0drp6nlXQpAop5PTUKNjdXMgtZVvrBL0JzZJebPP4,298
 parsl/executors/high_throughput/mpi_executor.py,sha256=khvGz56A8zU8XAY-R4TtqqiJB8B10mkVTXD_9xtrXgo,4696
 parsl/executors/high_throughput/mpi_prefix_composer.py,sha256=DmpKugANNa1bdYlqQBLHkrFc15fJpefPPhW9hkAlh1s,4308
 parsl/executors/high_throughput/mpi_resource_management.py,sha256=LFBbJ3BnzTcY_v-jNu30uoIB2Enk4cleN4ygY3dncjY,8194
 parsl/executors/high_throughput/probe.py,sha256=TNpGTXb4_DEeg_h-LHu4zEKi1-hffboxvKcZUl2OZGk,2751
-parsl/executors/high_throughput/process_worker_pool.py,sha256=3s-Ouo3ZEhod7hon8euyL37t1DbP5pSVjXyC23DSN_0,43075
+parsl/executors/high_throughput/process_worker_pool.py,sha256=ndV6uJBd7ErVRZdL9Iy1362m9y3k36zMSe8w3CM6eBg,43074
 parsl/executors/high_throughput/zmq_pipes.py,sha256=tAjQB3aNVMuTXziN3dbJWre46YpXgliD55qMBbhYTLU,8581
 parsl/executors/radical/__init__.py,sha256=CKbtV2numw5QvgIBq1htMUrt9TqDCIC2zifyf2svTNU,186
 parsl/executors/radical/executor.py,sha256=426cMt6d8uJFZ_7Ub1kCslaND4OKtBX5WZdz-0RXjMk,22554
@@ -176,7 +176,7 @@ parsl/providers/grid_engine/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMp
 parsl/providers/grid_engine/grid_engine.py,sha256=jTQjKaJh4eEXGbhrrCcXFV4AVFo2t39iVpslDR8gF6o,8565
 parsl/providers/grid_engine/template.py,sha256=a7iViKr8LXcFTPmsf_qQeVK5o_RekOAIlUOF0X1q-2M,273
 parsl/providers/kubernetes/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-parsl/providers/kubernetes/kube.py,sha256=umXdZqrpAIWh7OLsqZyF1L5AhmqmdpzKb3xTV3atKuk,14406
+parsl/providers/kubernetes/kube.py,sha256=ghOKM1gY1UjzMzWAheKsG15u8oUzRkXUyjtpqjkquIo,14952
 parsl/providers/kubernetes/template.py,sha256=VsRz6cmNaII-y4OdMT6sCwzQy95SJX6NMB0hmmFBhX4,50
 parsl/providers/local/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/providers/local/local.py,sha256=pTEcl9NnjRcL8FHcMeMEtJj1IXiAOxZ2Cih97Q5jDPY,11388
@@ -332,18 +332,20 @@ parsl/tests/test_error_handling/test_wrap_with_logs.py,sha256=aQDPECuhvZWUYQ6ysj
 parsl/tests/test_flowcontrol/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/tests/test_htex/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/tests/test_htex/test_basic.py,sha256=OCX4hbXLqxRakjt_pB9F68qJJv8qGOTkpiIzjHkSq1k,451
+parsl/tests/test_htex/test_block_manager_selector_unit.py,sha256=BeSj8jDeBHUEJVMVXwf0KLBhZ_pnsBEkG4vacldBfEY,737
 parsl/tests/test_htex/test_command_client_timeout.py,sha256=5tBViUhPT1ejnDDztTcEA690aA2BUxnPY0FpMf-1AXE,2008
 parsl/tests/test_htex/test_connected_blocks.py,sha256=gaXZSr__pIaLvKY6rF-4r1p_4dO5V28gtxHLT-psEFg,1640
 parsl/tests/test_htex/test_cpu_affinity_explicit.py,sha256=DVHrRCskDbJIrfB5YSi3ZSbfR4WzijA46aZfZzjNcrU,1382
 parsl/tests/test_htex/test_disconnected_blocks.py,sha256=3V1Ol9gMS6knjLTgIjB5GrunRSp4ANsJ_2vAvpyMR6c,1858
 parsl/tests/test_htex/test_disconnected_blocks_failing_provider.py,sha256=eOdipRpKMOkWAXB3UtY1UjqTiwfNs_csNLve8vllG_M,2040
-parsl/tests/test_htex/test_drain.py,sha256=Z2Z5-3NfLL9tMgJh4JkVKLZZDl3Z2gDAbEFHDSGdItw,2288
+parsl/tests/test_htex/test_drain.py,sha256=0MW3P71LUas2zmYFwwwat4G7dG8nVdxGeA3Fcy7NcMY,2454
 parsl/tests/test_htex/test_htex.py,sha256=J1uEGezic8ziPPZsQwfK9iNiTJ53NqXMhIg9CUunjZw,4901
 parsl/tests/test_htex/test_manager_failure.py,sha256=N-obuSZ8f7XA_XcddoN2LWKSVtpKUZvTHb7BFelS3iQ,1143
+parsl/tests/test_htex/test_manager_selector_by_block.py,sha256=D7EQVFeQw11w9XEgTeeCORp-HKE3kENpGGEMR9n609w,1386
 parsl/tests/test_htex/test_managers_command.py,sha256=Y-eUjtBzwW9erCYdph9bOesbkUvX8QUPqXt27DCgVS8,951
 parsl/tests/test_htex/test_missing_worker.py,sha256=gyp5i7_t-JHyJGtz_eXZKKBY5w8oqLOIxO6cJgGJMtQ,745
 parsl/tests/test_htex/test_multiple_disconnected_blocks.py,sha256=2vXZoIx4NuAWYuiNoL5Gxr85w72qZ7Kdb3JGh0FufTg,1867
-parsl/tests/test_htex/test_resource_spec_validation.py,sha256=JqboQRRFV0tEfWrGOdYT9pHazsUjyZLbF7qqnLFS_-A,914
+parsl/tests/test_htex/test_resource_spec_validation.py,sha256=VzOk4rjMNiDcEVLb-3YdlYZND7HRoGACJkTwq8NUTnc,1102
 parsl/tests/test_htex/test_worker_failure.py,sha256=Uz-RHI-LK78FMjXUvrUFmo4iYfmpDVBUcBxxRb3UG9M,603
 parsl/tests/test_htex/test_zmq_binding.py,sha256=Bq1HHuMxBE_AcaP1VZ-RqE4euCHO__Du05b2UZ5H1RA,3950
 parsl/tests/test_monitoring/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -365,6 +367,7 @@ parsl/tests/test_mpi_apps/test_mpiex.py,sha256=mlFdHK3A1B6NsEhxTQQX8lhs9qVza36FM
 parsl/tests/test_mpi_apps/test_resource_spec.py,sha256=5k6HM2jtb6sa7jetpI-Tl1nPQiN33VLaM7YT10c307E,3756
 parsl/tests/test_providers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/tests/test_providers/test_cobalt_deprecation_warning.py,sha256=UN2W6xJxuLx2euPqArORKFEU2VXez9_PYqq-0rZHanQ,391
+parsl/tests/test_providers/test_kubernetes_provider.py,sha256=AzCHfNz2HJwjP9BfxKH-XPaTHJCLXFErDMlQ_Ir8qRU,3861
 parsl/tests/test_providers/test_local_provider.py,sha256=R96E1eWgHVkvOQ1Au9wj-gfdWKAqGc-qlygFuxpGFQ8,7160
 parsl/tests/test_providers/test_pbspro_template.py,sha256=-bi1vags9yyNfpBxtjTqFjzMIg1VVPyf2M958UcXWmA,855
 parsl/tests/test_providers/test_slurm_instantiate.py,sha256=eW3pEZRIzZO1-eKFrBc7N5uoN5otwghgbqut74Kyqoc,500
@@ -455,6 +458,7 @@ parsl/tests/test_threads/test_configs.py,sha256=POwCKY4FMkI5-RSFz77cQYC7R_ISj0iA
 parsl/tests/test_threads/test_lazy_errors.py,sha256=6dJ65py5vUZkc0aRLL-LHuxBBp877Ktghb_jFDZsE98,543
 parsl/tests/test_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/tests/test_utils/test_representation_mixin.py,sha256=kUZeIDwA2rlbJ3-beGzLLwf3dOplTMCrWJN87etHcyY,1633
+parsl/tests/test_utils/test_sanitize_dns.py,sha256=8P_v5a5JLGU76OYf0LtclAwqJxGU0fMh_OZMVkMke3I,2954
 parsl/tests/unit/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parsl/tests/unit/test_file.py,sha256=vLycnYcv3bvSzL-FV8WdoibqTyb41BrH1LUYBavobsg,2850
 parsl/tests/unit/test_usage_tracking.py,sha256=xEfUlbBRpsFdUdOrCsk1Kz5AfmMxJT7f0_esZl8Ft-0,1884
@@ -462,13 +466,13 @@ parsl/usage_tracking/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hS
 parsl/usage_tracking/api.py,sha256=iaCY58Dc5J4UM7_dJzEEs871P1p1HdxBMtNGyVdzc9g,1821
 parsl/usage_tracking/levels.py,sha256=xbfzYEsd55KiZJ-mzNgPebvOH4rRHum04hROzEf41tU,291
 parsl/usage_tracking/usage.py,sha256=tcoZ2OUjsQVakG8Uu9_HFuEdzpSHyt4JarSRcLGnSMw,8918
-parsl-2024.10.14.data/scripts/exec_parsl_function.py,sha256=RUkJ4JSJAjr7YyRZ58zhMdg8cR5dVV9odUl3AuzNf3k,7802
-parsl-2024.10.14.data/scripts/interchange.py,sha256=FcEEmcuMcuFBB_aNOLzaYr5w3Yw9zKJxhtKbIUPVfhI,29912
-parsl-2024.10.14.data/scripts/parsl_coprocess.py,sha256=zrVjEqQvFOHxsLufPi00xzMONagjVwLZbavPM7bbjK4,5722
-parsl-2024.10.14.data/scripts/process_worker_pool.py,sha256=78QKnV5KbY_vcteC6k60gpDE4wEk6hsciet_qzs9QoU,43061
-parsl-2024.10.14.dist-info/LICENSE,sha256=tAkwu8-AdEyGxGoSvJ2gVmQdcicWw3j1ZZueVV74M-E,11357
-parsl-2024.10.14.dist-info/METADATA,sha256=55XMvDQssnRcrcG_9HLlmZnoesuPf61N6Kl5wKHlhnY,4072
-parsl-2024.10.14.dist-info/WHEEL,sha256=eOLhNAGa2EW3wWl_TU484h7q1UNgy0JXjjoqKoxAAQc,92
-parsl-2024.10.14.dist-info/entry_points.txt,sha256=XqnsWDYoEcLbsMcpnYGKLEnSBmaIe1YoM5YsBdJG2tI,176
-parsl-2024.10.14.dist-info/top_level.txt,sha256=PIheYoUFQtF2icLsgOykgU-Cjuwr2Oi6On2jo5RYgRM,6
-parsl-2024.10.14.dist-info/RECORD,,
+parsl-2024.10.21.data/scripts/exec_parsl_function.py,sha256=RUkJ4JSJAjr7YyRZ58zhMdg8cR5dVV9odUl3AuzNf3k,7802
+parsl-2024.10.21.data/scripts/interchange.py,sha256=FcEEmcuMcuFBB_aNOLzaYr5w3Yw9zKJxhtKbIUPVfhI,29912
+parsl-2024.10.21.data/scripts/parsl_coprocess.py,sha256=zrVjEqQvFOHxsLufPi00xzMONagjVwLZbavPM7bbjK4,5722
+parsl-2024.10.21.data/scripts/process_worker_pool.py,sha256=4K9vxwFHsz8QURwfq3VvnjEls7rYBxi2q0Gyy1cce5E,43060
+parsl-2024.10.21.dist-info/LICENSE,sha256=tAkwu8-AdEyGxGoSvJ2gVmQdcicWw3j1ZZueVV74M-E,11357
+parsl-2024.10.21.dist-info/METADATA,sha256=8hfXCgoISytZwjc6AefQ1vMFLcHWbaOAmBHOJNmg_Ds,4072
+parsl-2024.10.21.dist-info/WHEEL,sha256=eOLhNAGa2EW3wWl_TU484h7q1UNgy0JXjjoqKoxAAQc,92
+parsl-2024.10.21.dist-info/entry_points.txt,sha256=XqnsWDYoEcLbsMcpnYGKLEnSBmaIe1YoM5YsBdJG2tI,176
+parsl-2024.10.21.dist-info/top_level.txt,sha256=PIheYoUFQtF2icLsgOykgU-Cjuwr2Oi6On2jo5RYgRM,6
+parsl-2024.10.21.dist-info/RECORD,,