PyPI - skypilot-nightly - Versions diffs - 1.0.0.dev20250926__py3-none-any.whl → 1.0.0.dev20250927__py3-none-any.whl - Mend

skypilot-nightly 1.0.0.dev20250926py3-none-any.whl → 1.0.0.dev20250927py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.

This version of skypilot-nightly might be problematic. Click here for more details.

Files changed (57) hide show

sky/__init__.py +2 -2
sky/backends/backend_utils.py +25 -4
sky/backends/cloud_vm_ray_backend.py +151 -36
sky/client/cli/command.py +2 -1
sky/client/cli/table_utils.py +34 -0
sky/client/sdk.py +7 -5
sky/client/sdk_async.py +5 -5
sky/core.py +3 -4
sky/dashboard/out/404.html +1 -1
sky/dashboard/out/_next/static/{VXU6_xE28M55BOdwmUUJS → UDSEoDB67vwFMZyCJ4HWU}/_buildManifest.js +1 -1
sky/dashboard/out/_next/static/chunks/{3294.03e02ae73455f48e.js → 3294.93d9336bdc032b3a.js} +1 -1
sky/dashboard/out/_next/static/chunks/6856-5fdc9b851a18acdb.js +1 -0
sky/dashboard/out/_next/static/chunks/{webpack-8e64d11e58eab5cb.js → webpack-7340bc0f0dd8ae74.js} +1 -1
sky/dashboard/out/clusters/[cluster]/[job].html +1 -1
sky/dashboard/out/clusters/[cluster].html +1 -1
sky/dashboard/out/clusters.html +1 -1
sky/dashboard/out/config.html +1 -1
sky/dashboard/out/index.html +1 -1
sky/dashboard/out/infra/[context].html +1 -1
sky/dashboard/out/infra.html +1 -1
sky/dashboard/out/jobs/[job].html +1 -1
sky/dashboard/out/jobs/pools/[pool].html +1 -1
sky/dashboard/out/jobs.html +1 -1
sky/dashboard/out/users.html +1 -1
sky/dashboard/out/volumes.html +1 -1
sky/dashboard/out/workspace/new.html +1 -1
sky/dashboard/out/workspaces/[name].html +1 -1
sky/dashboard/out/workspaces.html +1 -1
sky/execution.py +0 -1
sky/global_user_state.py +3 -3
sky/jobs/server/core.py +96 -26
sky/jobs/server/utils.py +65 -32
sky/jobs/state.py +145 -3
sky/jobs/utils.py +85 -7
sky/schemas/api/responses.py +18 -0
sky/schemas/generated/managed_jobsv1_pb2.py +70 -0
sky/schemas/generated/managed_jobsv1_pb2.pyi +262 -0
sky/schemas/generated/managed_jobsv1_pb2_grpc.py +278 -0
sky/serve/serve_utils.py +16 -0
sky/serve/server/core.py +1 -1
sky/serve/server/impl.py +6 -6
sky/server/requests/serializers/decoders.py +2 -2
sky/server/requests/serializers/encoders.py +7 -3
sky/skylet/constants.py +1 -1
sky/skylet/job_lib.py +2 -32
sky/skylet/log_lib.py +211 -0
sky/skylet/log_lib.pyi +30 -1
sky/skylet/services.py +208 -2
sky/skylet/skylet.py +3 -0
{skypilot_nightly-1.0.0.dev20250926.dist-info → skypilot_nightly-1.0.0.dev20250927.dist-info}/METADATA +32 -32
{skypilot_nightly-1.0.0.dev20250926.dist-info → skypilot_nightly-1.0.0.dev20250927.dist-info}/RECORD +56 -52
sky/dashboard/out/_next/static/chunks/6856-2b3600ff2854d066.js +0 -1
/sky/dashboard/out/_next/static/{VXU6_xE28M55BOdwmUUJS → UDSEoDB67vwFMZyCJ4HWU}/_ssgManifest.js +0 -0
{skypilot_nightly-1.0.0.dev20250926.dist-info → skypilot_nightly-1.0.0.dev20250927.dist-info}/WHEEL +0 -0
{skypilot_nightly-1.0.0.dev20250926.dist-info → skypilot_nightly-1.0.0.dev20250927.dist-info}/entry_points.txt +0 -0
{skypilot_nightly-1.0.0.dev20250926.dist-info → skypilot_nightly-1.0.0.dev20250927.dist-info}/licenses/LICENSE +0 -0
{skypilot_nightly-1.0.0.dev20250926.dist-info → skypilot_nightly-1.0.0.dev20250927.dist-info}/top_level.txt +0 -0

sky/dashboard/out/workspaces.html CHANGED Viewed

	@@ -1 +1 @@
1	- <!DOCTYPE html><html><head><meta charSet="utf-8"/><meta name="viewport" content="width=device-width"/><meta name="next-head-count" content="2"/><link rel="preload" href="/dashboard/_next/static/css/4614e06482d7309e.css" as="style"/><link rel="stylesheet" href="/dashboard/_next/static/css/4614e06482d7309e.css" data-n-g=""/><noscript data-n-css=""></noscript><script defer="" nomodule="" src="/dashboard/_next/static/chunks/polyfills-78c92fac7aa8fdd8.js"></script><script src="/dashboard/_next/static/chunks/webpack-~~8e64d11e58eab5cb~~.js" defer=""></script><script src="/dashboard/_next/static/chunks/framework-cf60a09ccd051a10.js" defer=""></script><script src="/dashboard/_next/static/chunks/main-f15ccb73239a3bf1.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/_app-ce361c6959bc2001.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/workspaces-7528cc0ef8c522c5.js" defer=""></script><script src="/dashboard/_next/static/~~VXU6_xE28M55BOdwmUUJS~~/_buildManifest.js" defer=""></script><script src="/dashboard/_next/static/~~VXU6_xE28M55BOdwmUUJS~~/_ssgManifest.js" defer=""></script></head><body><div id="__next"></div><script id="__NEXT_DATA__" type="application/json">{"props":{"pageProps":{}},"page":"/workspaces","query":{},"buildId":"~~VXU6_xE28M55BOdwmUUJS~~","assetPrefix":"/dashboard","nextExport":true,"autoExport":true,"isFallback":false,"scriptLoader":[]}</script></body></html>
1	+ <!DOCTYPE html><html><head><meta charSet="utf-8"/><meta name="viewport" content="width=device-width"/><meta name="next-head-count" content="2"/><link rel="preload" href="/dashboard/_next/static/css/4614e06482d7309e.css" as="style"/><link rel="stylesheet" href="/dashboard/_next/static/css/4614e06482d7309e.css" data-n-g=""/><noscript data-n-css=""></noscript><script defer="" nomodule="" src="/dashboard/_next/static/chunks/polyfills-78c92fac7aa8fdd8.js"></script><script src="/dashboard/_next/static/chunks/webpack-7340bc0f0dd8ae74.js" defer=""></script><script src="/dashboard/_next/static/chunks/framework-cf60a09ccd051a10.js" defer=""></script><script src="/dashboard/_next/static/chunks/main-f15ccb73239a3bf1.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/_app-ce361c6959bc2001.js" defer=""></script><script src="/dashboard/_next/static/chunks/pages/workspaces-7528cc0ef8c522c5.js" defer=""></script><script src="/dashboard/_next/static/UDSEoDB67vwFMZyCJ4HWU/_buildManifest.js" defer=""></script><script src="/dashboard/_next/static/UDSEoDB67vwFMZyCJ4HWU/_ssgManifest.js" defer=""></script></head><body><div id="__next"></div><script id="__NEXT_DATA__" type="application/json">{"props":{"pageProps":{}},"page":"/workspaces","query":{},"buildId":"UDSEoDB67vwFMZyCJ4HWU","assetPrefix":"/dashboard","nextExport":true,"autoExport":true,"isFallback":false,"scriptLoader":[]}</script></body></html>

sky/execution.py CHANGED Viewed

@@ -673,7 +673,6 @@ def launch(
     # see the setup logs when inspecting the launch process to know
     # excatly what the job is waiting for.
     detach_setup = controller_utils.Controllers.from_name(cluster_name) is None
     return _execute(
         entrypoint=entrypoint,
         dryrun=dryrun,

sky/global_user_state.py CHANGED Viewed

@@ -483,7 +483,7 @@ def get_user(user_id: str) -> Optional[models.User]:
 @_init_db
 @metrics_lib.time_me
-def _get_users(user_ids: Set[str]) -> Dict[str, models.User]:
+def get_users(user_ids: Set[str]) -> Dict[str, models.User]:
     assert _SQLALCHEMY_ENGINE is not None
     with orm.Session(_SQLALCHEMY_ENGINE) as session:
         rows = session.query(user_table).filter(
@@ -1659,7 +1659,7 @@ def get_clusters(
     # get all users needed for the rows at once
     user_hashes = set(row_to_user_hash.values())
-    user_hash_to_user = _get_users(user_hashes)
+    user_hash_to_user = get_users(user_hashes)
     # get last cluster event for each row
     cluster_hashes = set(row_to_user_hash.keys())
@@ -1807,7 +1807,7 @@ def get_clusters_from_history(
         row_to_user_hash[row.cluster_hash] = user_hash
     user_hashes = set(row_to_user_hash.values())
-    user_hash_to_user = _get_users(user_hashes)
+    user_hash_to_user = get_users(user_hashes)
     cluster_hashes = set(row_to_user_hash.keys())
     if not abbreviate_response:
         last_cluster_event_dict = _get_last_cluster_event_multiple(

sky/jobs/server/core.py CHANGED Viewed

@@ -19,7 +19,9 @@ from sky import provision as provision_lib
 from sky import sky_logging
 from sky import skypilot_config
 from sky import task as task_lib
+from sky.adaptors import common as adaptors_common
 from sky.backends import backend_utils
+from sky.backends import cloud_vm_ray_backend
 from sky.catalog import common as service_catalog_common
 from sky.data import storage as storage_lib
 from sky.jobs import constants as managed_job_constants
@@ -44,8 +46,15 @@ from sky.utils import ux_utils
 from sky.workspaces import core as workspaces_core
 if typing.TYPE_CHECKING:
+    from google.protobuf import json_format
     import sky
-    from sky.backends import cloud_vm_ray_backend
+    from sky.schemas.generated import managed_jobsv1_pb2
+else:
+    json_format = adaptors_common.LazyImport('google.protobuf.json_format')
+    managed_jobsv1_pb2 = adaptors_common.LazyImport(
+        'sky.schemas.generated.managed_jobsv1_pb2')
 logger = sky_logging.init_logger(__name__)
@@ -701,11 +710,13 @@ def queue_v2(
     assert isinstance(backend, backends.CloudVmRayBackend)
     user_hashes: Optional[List[Optional[str]]] = None
+    show_jobs_without_user_hash = False
     if not all_users:
         user_hashes = [common_utils.get_user_hash()]
         # For backwards compatibility, we show jobs that do not have a
         # user_hash. TODO(cooperc): Remove before 0.12.0.
         user_hashes.append(None)
+        show_jobs_without_user_hash = True
     elif user_match is not None:
         users = global_user_state.get_user_by_name_match(user_match)
         if not users:
@@ -713,6 +724,38 @@ def queue_v2(
         user_hashes = [user.id for user in users]
     accessible_workspaces = list(workspaces_core.get_workspaces().keys())
+    if handle.is_grpc_enabled_with_flag:
+        try:
+            request = managed_jobsv1_pb2.GetJobTableRequest(
+                skip_finished=skip_finished,
+                accessible_workspaces=accessible_workspaces,
+                job_ids=managed_jobsv1_pb2.JobIds(
+                    ids=job_ids) if job_ids is not None else None,
+                workspace_match=workspace_match,
+                name_match=name_match,
+                pool_match=pool_match,
+                page=page,
+                limit=limit,
+                # Remove None from user_hashes, as the gRPC server uses the
+                # show_jobs_without_user_hash flag instead.
+                user_hashes=managed_jobsv1_pb2.UserHashes(hashes=[
+                    user_hash for user_hash in user_hashes
+                    if user_hash is not None
+                ]) if user_hashes is not None else None,
+                statuses=managed_jobsv1_pb2.Statuses(
+                    statuses=statuses) if statuses is not None else None,
+                show_jobs_without_user_hash=show_jobs_without_user_hash,
+            )
+            response = backend_utils.invoke_skylet_with_retries(
+                lambda: cloud_vm_ray_backend.SkyletClient(
+                    handle.get_grpc_channel()).get_managed_job_table(request))
+            jobs = managed_job_utils.decode_managed_job_protos(response.jobs)
+            return jobs, response.total, dict(
+                response.status_counts), response.total_no_filter
+        except exceptions.SkyletMethodNotImplementedError:
+            pass
     code = managed_job_utils.ManagedJobCodeGen.get_job_table(
         skip_finished, accessible_workspaces, job_ids, workspace_match,
         name_match, pool_match, page, limit, user_hashes, statuses)
@@ -819,33 +862,60 @@ def cancel(name: Optional[str] = None,
                     'Can only specify one of JOB_IDS, name, pool, or all/'
                     f'all_users. Provided {" ".join(arguments)!r}.')
+        job_ids = None if (all_users or all) else job_ids
         backend = backend_utils.get_backend_from_handle(handle)
         assert isinstance(backend, backends.CloudVmRayBackend)
-        if all_users:
-            code = managed_job_utils.ManagedJobCodeGen.cancel_jobs_by_id(
-                None, all_users=True)
-        elif all:
-            code = managed_job_utils.ManagedJobCodeGen.cancel_jobs_by_id(None)
-        elif job_ids:
-            code = managed_job_utils.ManagedJobCodeGen.cancel_jobs_by_id(
-                job_ids)
-        elif name is not None:
-            code = managed_job_utils.ManagedJobCodeGen.cancel_job_by_name(name)
-        else:
-            assert pool is not None, (job_ids, name, pool, all)
-            code = managed_job_utils.ManagedJobCodeGen.cancel_jobs_by_pool(pool)
-        # The stderr is redirected to stdout
-        returncode, stdout, stderr = backend.run_on_head(handle,
-                                                         code,
-                                                         require_outputs=True,
-                                                         stream_logs=False)
-        try:
-            subprocess_utils.handle_returncode(returncode, code,
-                                               'Failed to cancel managed job',
-                                               stdout + stderr)
-        except exceptions.CommandError as e:
-            with ux_utils.print_exception_no_traceback():
-                raise RuntimeError(e.error_msg) from e
+        use_legacy = not handle.is_grpc_enabled_with_flag
+        if not use_legacy:
+            current_workspace = skypilot_config.get_active_workspace()
+            try:
+                request = managed_jobsv1_pb2.CancelJobsRequest(
+                    current_workspace=current_workspace)
+                if all_users or all or job_ids:
+                    request.all_users = all_users
+                    if all:
+                        request.user_hash = common_utils.get_user_hash()
+                    if job_ids is not None:
+                        request.job_ids.CopyFrom(
+                            managed_jobsv1_pb2.JobIds(ids=job_ids))
+                elif name is not None:
+                    request.job_name = name
+                else:
+                    assert pool is not None, (job_ids, name, pool, all)
+                    request.pool_name = pool
+                response = backend_utils.invoke_skylet_with_retries(
+                    lambda: cloud_vm_ray_backend.SkyletClient(
+                        handle.get_grpc_channel()).cancel_managed_jobs(request))
+                stdout = response.message
+            except exceptions.SkyletMethodNotImplementedError:
+                use_legacy = True
+        if use_legacy:
+            if all_users or all or job_ids:
+                code = managed_job_utils.ManagedJobCodeGen.cancel_jobs_by_id(
+                    job_ids, all_users=all_users)
+            elif name is not None:
+                code = managed_job_utils.ManagedJobCodeGen.cancel_job_by_name(
+                    name)
+            else:
+                assert pool is not None, (job_ids, name, pool, all)
+                code = managed_job_utils.ManagedJobCodeGen.cancel_jobs_by_pool(
+                    pool)
+            # The stderr is redirected to stdout
+            returncode, stdout, stderr = backend.run_on_head(
+                handle, code, require_outputs=True, stream_logs=False)
+            try:
+                subprocess_utils.handle_returncode(
+                    returncode, code, 'Failed to cancel managed job',
+                    stdout + stderr)
+            except exceptions.CommandError as e:
+                with ux_utils.print_exception_no_traceback():
+                    raise RuntimeError(e.error_msg) from e
         logger.info(stdout)
         if 'Multiple jobs found with name' in stdout:

sky/jobs/server/utils.py CHANGED Viewed

@@ -1,13 +1,24 @@
 """Utility functions for managed jobs."""
+import typing
 from sky import backends
+from sky import exceptions
 from sky import sky_logging
+from sky.adaptors import common as adaptors_common
 from sky.backends import backend_utils
+from sky.backends import cloud_vm_ray_backend
 from sky.jobs import utils as managed_job_utils
 from sky.skylet import constants as skylet_constants
 from sky.utils import controller_utils
 logger = sky_logging.init_logger(__name__)
+if typing.TYPE_CHECKING:
+    from sky.schemas.generated import managed_jobsv1_pb2
+else:
+    managed_jobsv1_pb2 = adaptors_common.LazyImport(
+        'sky.schemas.generated.managed_jobsv1_pb2')
 def check_version_mismatch_and_non_terminal_jobs() -> None:
     """Check if controller has version mismatch and non-terminal jobs exist.
@@ -28,42 +39,64 @@ def check_version_mismatch_and_non_terminal_jobs() -> None:
     backend = backend_utils.get_backend_from_handle(handle)
     assert isinstance(backend, backends.CloudVmRayBackend)
-    # Get controller version and raw job table
-    code = managed_job_utils.ManagedJobCodeGen.get_version_and_job_table()
-    returncode, output, stderr = backend.run_on_head(handle,
-                                                     code,
-                                                     require_outputs=True,
-                                                     stream_logs=False,
-                                                     separate_stderr=True)
-    if returncode != 0:
-        logger.error(output + stderr)
-        raise ValueError('Failed to check controller version and jobs with '
-                         f'returncode: {returncode}.\n{output + stderr}')
-    # Parse the output to extract controller version (split only on first
-    # newline)
-    output_parts = output.strip().split('\n', 1)
-    # Extract controller version from first line
-    if len(output_parts) < 2 or not output_parts[0].startswith(
-            'controller_version:'):
-        raise ValueError(
-            f'Expected controller version in first line, got: {output}')
-    controller_version = output_parts[0].split(':', 1)[1]
-    # Rest is job table payload (preserving any newlines within it)
-    job_table_payload = output_parts[1]
+    use_legacy = not handle.is_grpc_enabled_with_flag
+    if not use_legacy:
+        try:
+            version_request = managed_jobsv1_pb2.GetVersionRequest()
+            version_response = backend_utils.invoke_skylet_with_retries(
+                lambda: cloud_vm_ray_backend.SkyletClient(
+                    handle.get_grpc_channel(
+                    )).get_managed_job_controller_version(version_request))
+            controller_version = version_response.controller_version
+            job_table_request = managed_jobsv1_pb2.GetJobTableRequest()
+            job_table_response = backend_utils.invoke_skylet_with_retries(
+                lambda: cloud_vm_ray_backend.SkyletClient(
+                    handle.get_grpc_channel()).get_managed_job_table(
+                        job_table_request))
+            jobs = managed_job_utils.decode_managed_job_protos(
+                job_table_response.jobs)
+        except exceptions.SkyletMethodNotImplementedError:
+            use_legacy = True
+    if use_legacy:
+        # Get controller version and raw job table
+        code = managed_job_utils.ManagedJobCodeGen.get_version_and_job_table()
+        returncode, output, stderr = backend.run_on_head(handle,
+                                                         code,
+                                                         require_outputs=True,
+                                                         stream_logs=False,
+                                                         separate_stderr=True)
+        if returncode != 0:
+            logger.error(output + stderr)
+            raise ValueError('Failed to check controller version and jobs with '
+                             f'returncode: {returncode}.\n{output + stderr}')
+        # Parse the output to extract controller version (split only on first
+        # newline)
+        output_parts = output.strip().split('\n', 1)
+        # Extract controller version from first line
+        if len(output_parts) < 2 or not output_parts[0].startswith(
+                'controller_version:'):
+            raise ValueError(
+                f'Expected controller version in first line, got: {output}')
+        controller_version = output_parts[0].split(':', 1)[1]
+        # Rest is job table payload (preserving any newlines within it)
+        job_table_payload = output_parts[1]
+        # Load and filter jobs locally using existing method
+        jobs, _, _, _, _ = managed_job_utils.load_managed_job_queue(
+            job_table_payload)
     # Process locally: check version match and filter non-terminal jobs
     version_matches = (controller_version == local_version or
                        int(controller_version) > 17)
-    # Load and filter jobs locally using existing method
-    jobs, _, _, _, _ = managed_job_utils.load_managed_job_queue(
-        job_table_payload)
     non_terminal_jobs = [job for job in jobs if not job['status'].is_terminal()]
     has_non_terminal_jobs = len(non_terminal_jobs) > 0

sky/jobs/state.py CHANGED Viewed

@@ -25,6 +25,7 @@ from sqlalchemy.ext import declarative
 from sky import exceptions
 from sky import sky_logging
 from sky import skypilot_config
+from sky.adaptors import common as adaptors_common
 from sky.skylet import constants
 from sky.utils import common_utils
 from sky.utils import context_utils
@@ -34,6 +35,11 @@ from sky.utils.db import migration_utils
 if typing.TYPE_CHECKING:
     from sqlalchemy.engine import row
+    from sky.schemas.generated import managed_jobsv1_pb2
+else:
+    managed_jobsv1_pb2 = adaptors_common.LazyImport(
+        'sky.schemas.generated.managed_jobsv1_pb2')
 # Separate callback types for sync and async contexts
 SyncCallbackType = Callable[[str], None]
 AsyncCallbackType = Callable[[str], Awaitable[Any]]
@@ -448,6 +454,75 @@ class ManagedJobStatus(enum.Enum):
             cls.RECOVERING,
         ]
+    @classmethod
+    def from_protobuf(
+        cls, protobuf_value: 'managed_jobsv1_pb2.ManagedJobStatus'
+    ) -> Optional['ManagedJobStatus']:
+        """Convert protobuf ManagedJobStatus enum to Python enum value."""
+        protobuf_to_enum = {
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_UNSPECIFIED: None,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_PENDING: cls.PENDING,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_SUBMITTED:
+                cls.DEPRECATED_SUBMITTED,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_STARTING: cls.STARTING,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_RUNNING: cls.RUNNING,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_SUCCEEDED: cls.SUCCEEDED,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED: cls.FAILED,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_CONTROLLER:
+                cls.FAILED_CONTROLLER,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_SETUP:
+                cls.FAILED_SETUP,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_CANCELLED: cls.CANCELLED,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_RECOVERING: cls.RECOVERING,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_CANCELLING: cls.CANCELLING,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_PRECHECKS:
+                cls.FAILED_PRECHECKS,
+            managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_NO_RESOURCE:
+                cls.FAILED_NO_RESOURCE,
+        }
+        if protobuf_value not in protobuf_to_enum:
+            raise ValueError(
+                f'Unknown protobuf ManagedJobStatus value: {protobuf_value}')
+        return protobuf_to_enum[protobuf_value]
+    def to_protobuf(self) -> 'managed_jobsv1_pb2.ManagedJobStatus':
+        """Convert this Python enum value to protobuf enum value."""
+        enum_to_protobuf = {
+            ManagedJobStatus.PENDING:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_PENDING,
+            ManagedJobStatus.DEPRECATED_SUBMITTED:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_SUBMITTED,
+            ManagedJobStatus.STARTING:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_STARTING,
+            ManagedJobStatus.RUNNING:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_RUNNING,
+            ManagedJobStatus.SUCCEEDED:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_SUCCEEDED,
+            ManagedJobStatus.FAILED:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED,
+            ManagedJobStatus.FAILED_CONTROLLER:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_CONTROLLER,
+            ManagedJobStatus.FAILED_SETUP:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_SETUP,
+            ManagedJobStatus.CANCELLED:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_CANCELLED,
+            ManagedJobStatus.RECOVERING:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_RECOVERING,
+            ManagedJobStatus.CANCELLING:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_CANCELLING,
+            ManagedJobStatus.FAILED_PRECHECKS:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_PRECHECKS,
+            ManagedJobStatus.FAILED_NO_RESOURCE:
+                managed_jobsv1_pb2.MANAGED_JOB_STATUS_FAILED_NO_RESOURCE,
+        }
+        if self not in enum_to_protobuf:
+            raise ValueError(f'Unknown ManagedJobStatus value: {self}')
+        return enum_to_protobuf[self]
 _SPOT_STATUS_TO_COLOR = {
     ManagedJobStatus.PENDING: colorama.Fore.BLUE,
@@ -537,6 +612,60 @@ class ManagedJobScheduleState(enum.Enum):
     # The job is in a terminal state. (Not necessarily SUCCEEDED.)
     DONE = 'DONE'
+    @classmethod
+    def from_protobuf(
+        cls, protobuf_value: 'managed_jobsv1_pb2.ManagedJobScheduleState'
+    ) -> Optional['ManagedJobScheduleState']:
+        """Convert protobuf ManagedJobScheduleState enum to Python enum value.
+        """
+        protobuf_to_enum = {
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_UNSPECIFIED: None,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_INVALID: cls.INVALID,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_INACTIVE:
+                cls.INACTIVE,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_WAITING: cls.WAITING,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_ALIVE_WAITING:
+                cls.ALIVE_WAITING,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_LAUNCHING:
+                cls.LAUNCHING,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_ALIVE_BACKOFF:
+                cls.ALIVE_BACKOFF,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_ALIVE: cls.ALIVE,
+            managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_DONE: cls.DONE,
+        }
+        if protobuf_value not in protobuf_to_enum:
+            raise ValueError('Unknown protobuf ManagedJobScheduleState value: '
+                             f'{protobuf_value}')
+        return protobuf_to_enum[protobuf_value]
+    def to_protobuf(self) -> 'managed_jobsv1_pb2.ManagedJobScheduleState':
+        """Convert this Python enum value to protobuf enum value."""
+        enum_to_protobuf = {
+            ManagedJobScheduleState.INVALID:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_INVALID,
+            ManagedJobScheduleState.INACTIVE:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_INACTIVE,
+            ManagedJobScheduleState.WAITING:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_WAITING,
+            ManagedJobScheduleState.ALIVE_WAITING:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_ALIVE_WAITING,
+            ManagedJobScheduleState.LAUNCHING:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_LAUNCHING,
+            ManagedJobScheduleState.ALIVE_BACKOFF:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_ALIVE_BACKOFF,
+            ManagedJobScheduleState.ALIVE:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_ALIVE,
+            ManagedJobScheduleState.DONE:
+                managed_jobsv1_pb2.MANAGED_JOB_SCHEDULE_STATE_DONE,
+        }
+        if self not in enum_to_protobuf:
+            raise ValueError(f'Unknown ManagedJobScheduleState value: {self}')
+        return enum_to_protobuf[self]
 # === Status transition functions ===
 @_init_db
@@ -792,8 +921,14 @@ def set_local_log_file(job_id: int, task_id: Optional[int],
 # ======== utility functions ========
 @_init_db
 def get_nonterminal_job_ids_by_name(name: Optional[str],
+                                    user_hash: Optional[str] = None,
                                     all_users: bool = False) -> List[int]:
-    """Get non-terminal job ids by name."""
+    """Get non-terminal job ids by name.
+    If name is None:
+    1. if all_users is False, get for the given user_hash
+    2. otherwise, get for all users
+    """
     assert _SQLALCHEMY_ENGINE is not None
     with orm.Session(_SQLALCHEMY_ENGINE) as session:
@@ -810,8 +945,15 @@ def get_nonterminal_job_ids_by_name(name: Optional[str],
             ])
         ]
         if name is None and not all_users:
-            where_conditions.append(
-                job_info_table.c.user_hash == common_utils.get_user_hash())
+            if user_hash is None:
+                # For backwards compatibility. With codegen, USER_ID_ENV_VAR
+                # was set to the correct value by the jobs controller, as
+                # part of ManagedJobCodeGen._build(). This is no longer the
+                # case for the Skylet gRPC server, which is why we need to
+                # pass it explicitly through the request body.
+                logger.debug('user_hash is None, using current user hash')
+                user_hash = common_utils.get_user_hash()
+            where_conditions.append(job_info_table.c.user_hash == user_hash)
         if name is not None:
             # We match the job name from `job_info` for the jobs submitted after
             # #1982, and from `spot` for the jobs submitted before #1982, whose

skypilot-nightly 1.0.0.dev20250926__py3-none-any.whl → 1.0.0.dev20250927__py3-none-any.whl

Potentially problematic release.

skypilot-nightly 1.0.0.dev20250926py3-none-any.whl → 1.0.0.dev20250927py3-none-any.whl