[core] (cgroups 14/n) Clean up bazel targets and support cross-platform build. #57244

israbbani · 2025-10-06T23:14:20Z

For more details about the resource isolation project see #54703.

This PR introduces two public bazel targets from the //src/ray/common/cgroup2 subsystem.

CgroupManagerFactory is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux.
CgroupManagerInterface is the public API of CgroupManager.

It also introduces a few other changes

All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory.
NodeManager now controls the lifecycle (and destruction) of CgroupManager.
SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files.

CgroupManagerFactory which constructs a cross-platform cgroup manager with selective compilation Signed-off-by: irabbani <[email protected]>

and CgroupManagerFactory are the only public targets. CgroupManagerFactory will delegate to the appropriate implementation for each platform. Signed-off-by: irabbani <[email protected]>

build. The Cgroup subsystem only exposes CgroupManagerInterface and CgroupManagerFactory as public targets. Signed-off-by: irabbani <[email protected]>

Signed-off-by: irabbani <[email protected]>

… irabbani/cgroups-14 Signed-off-by: irabbani <[email protected]>

Signed-off-by: irabbani <[email protected]>

… irabbani/cgroups-14 Signed-off-by: irabbani <[email protected]>

Signed-off-by: irabbani <[email protected]>

Signed-off-by: israbbani <[email protected]>

Signed-off-by: irabbani <[email protected]>

… irabbani/cgroups-14 Signed-off-by: irabbani <[email protected]>

edoakes

Looks great!

edoakes · 2025-10-07T20:57:00Z

src/ray/common/cgroup2/cgroup_manager_factory.h

+// TODO(54703): Refactor the configs into a struct called CgroupManagerConfig
+// and delegate input validation and error messages to it.


src/ray/common/cgroup2/cgroup_manager_factory.h

edoakes · 2025-10-07T20:58:43Z

src/ray/common/cgroup2/linux_cgroup_manager_factory.cc

+  RAY_CHECK(!cgroup_path.empty())
+      << "Failed to start CgroupManager. If enable_resource_isolation is set to true, "
+         "cgroup_path cannot be empty.";


in the general case, should we structure the factories so that they RAY_CHECK on initialization failure or return a status code? the paranoid/perfectionist version would be the latter. practically I'm not sure that it makes much of a difference.

Capturing our discussion for posterity.

If there's a panic or a fatal error, I would keep the actual RAY_CHECK as close to the error as possible. You're not worried about clean up and you can dump as much context as possible.

I would return a Status is if there's a chance that the caller can either recover provide more useful information.
However, this has the unpleasant side-effect of having landmines buried deep in the call stack.. which makes me sad.

The antidote to this sadness is to have as few FATAL errors as possible and most failures should be recoverable.

Invariant checking on component startups or detecting misconfigurations is a valid use of RAY_CHECK.

src/ray/common/cgroup2/sysfs_cgroup_driver.h

edoakes · 2025-10-07T21:00:49Z

minor comments; ping when ready to merge

Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: Ibrahim Rabbani <[email protected]>

Signed-off-by: irabbani <[email protected]>

israbbani · 2025-10-07T23:10:33Z

Gonna trigger 1 more set of MacOS and Windows tests. I'll let you know when they pass.

Signed-off-by: israbbani <[email protected]>

israbbani · 2025-10-09T15:20:05Z

@edoakes the MacOS failure looks unrelated to the change or to core:

(11:07:33) INFO: Running command line: bazel-bin/ci/ray_ci/automation/test_db_bot core /tmp/bazel_event_logs
--
  | Traceback (most recent call last):
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/io_ray/ci/ray_ci/automation/test_db_bot.py", line 41, in <module>
  | main()
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/py_deps_buildkite_click/click/core.py", line 1157, in __call__
  | return self.main(*args, **kwargs)
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/py_deps_buildkite_click/click/core.py", line 1078, in main
  | rv = self.invoke(ctx)
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/py_deps_buildkite_click/click/core.py", line 1434, in invoke
  | return ctx.invoke(self.callback, **ctx.params)
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/py_deps_buildkite_click/click/core.py", line 783, in invoke
  | return __callback(*args, **kwargs)
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/io_ray/ci/ray_ci/automation/test_db_bot.py", line 36, in main
  | TesterContainer.upload_test_results(team, bazel_log_dir)
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/io_ray/ci/ray_ci/tester_container.py", line 158, in upload_test_results
  | for test, result in cls.get_test_and_results(team, bazel_log_dir):
  | File "/private/var/tmp/_bazel_ec2-user/b625f2b2abd7a5bdbf5aca7c03f4a916/execroot/io_ray/bazel-out/darwin-opt/bin/ci/ray_ci/automation/test_db_bot.runfiles/io_ray/ci/ray_ci/tester_container.py", line 191, in get_test_and_results
  | for file in listdir(bazel_log_dir):
  | FileNotFoundError: [Errno 2] No such file or directory: '/tmp/bazel_event_logs'```

…rm build. (ray-project#57244) For more details about the resource isolation project see ray-project#54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: Josh Kodi <[email protected]>

…ses. (#57269) This PR stacks on #57244. For more details about the resource isolation project see #54703. In the previous ray cgroup hierarchy, all processes that were in the path `--cgroup-path` were moved into the system cgroup. This changes the hierarchy to now have a separate cgroup for all non-ray processes. The new cgroup hierarchy looks like ``` cgroup_path (e.g. /sys/fs/cgroup) | ray-node_<node_id> | | system user | | | leaf workers non-ray ``` The cgroups contain the following processes * system/leaf (all ray non-worker processes e.g. raylet, runtime_env_agent, gcs_server, ...) * user/workers (all ray worker processes) * user/non-ray (all non-ray processes migrated from cgroup_path). Note: If you're running ray inside a container, all non-ray processes running in the container will be migrated to `user/non-ray` The following controllers will be enabled * cgroup_path (cpu, memory) * ray-node_<node_id> (cpu, memory) * system (memory) The following constraints are applied * system (cpu.weight, memory.min) * user (cpu.weight) --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…rm build. (ray-project#57244) For more details about the resource isolation project see ray-project#54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…rm build. (#57244) For more details about the resource isolation project see #54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…ses. (#57269) This PR stacks on #57244. For more details about the resource isolation project see #54703. In the previous ray cgroup hierarchy, all processes that were in the path `--cgroup-path` were moved into the system cgroup. This changes the hierarchy to now have a separate cgroup for all non-ray processes. The new cgroup hierarchy looks like ``` cgroup_path (e.g. /sys/fs/cgroup) | ray-node_<node_id> | | system user | | | leaf workers non-ray ``` The cgroups contain the following processes * system/leaf (all ray non-worker processes e.g. raylet, runtime_env_agent, gcs_server, ...) * user/workers (all ray worker processes) * user/non-ray (all non-ray processes migrated from cgroup_path). Note: If you're running ray inside a container, all non-ray processes running in the container will be migrated to `user/non-ray` The following controllers will be enabled * cgroup_path (cpu, memory) * ray-node_<node_id> (cpu, memory) * system (memory) The following constraints are applied * system (cpu.weight, memory.min) * user (cpu.weight) --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…rm build. (ray-project#57244) For more details about the resource isolation project see ray-project#54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…ses. (ray-project#57269) This PR stacks on ray-project#57244. For more details about the resource isolation project see ray-project#54703. In the previous ray cgroup hierarchy, all processes that were in the path `--cgroup-path` were moved into the system cgroup. This changes the hierarchy to now have a separate cgroup for all non-ray processes. The new cgroup hierarchy looks like ``` cgroup_path (e.g. /sys/fs/cgroup) | ray-node_<node_id> | | system user | | | leaf workers non-ray ``` The cgroups contain the following processes * system/leaf (all ray non-worker processes e.g. raylet, runtime_env_agent, gcs_server, ...) * user/workers (all ray worker processes) * user/non-ray (all non-ray processes migrated from cgroup_path). Note: If you're running ray inside a container, all non-ray processes running in the container will be migrated to `user/non-ray` The following controllers will be enabled * cgroup_path (cpu, memory) * ray-node_<node_id> (cpu, memory) * system (memory) The following constraints are applied * system (cpu.weight, memory.min) * user (cpu.weight) --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…rm build. (ray-project#57244) For more details about the resource isolation project see ray-project#54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: xgui <[email protected]>

…ses. (ray-project#57269) This PR stacks on ray-project#57244. For more details about the resource isolation project see ray-project#54703. In the previous ray cgroup hierarchy, all processes that were in the path `--cgroup-path` were moved into the system cgroup. This changes the hierarchy to now have a separate cgroup for all non-ray processes. The new cgroup hierarchy looks like ``` cgroup_path (e.g. /sys/fs/cgroup) | ray-node_<node_id> | | system user | | | leaf workers non-ray ``` The cgroups contain the following processes * system/leaf (all ray non-worker processes e.g. raylet, runtime_env_agent, gcs_server, ...) * user/workers (all ray worker processes) * user/non-ray (all non-ray processes migrated from cgroup_path). Note: If you're running ray inside a container, all non-ray processes running in the container will be migrated to `user/non-ray` The following controllers will be enabled * cgroup_path (cpu, memory) * ray-node_<node_id> (cpu, memory) * system (memory) The following constraints are applied * system (cpu.weight, memory.min) * user (cpu.weight) --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: xgui <[email protected]>

…rm build. (#57244) For more details about the resource isolation project see #54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: elliot-barn <[email protected]>

…ses. (#57269) This PR stacks on #57244. For more details about the resource isolation project see #54703. In the previous ray cgroup hierarchy, all processes that were in the path `--cgroup-path` were moved into the system cgroup. This changes the hierarchy to now have a separate cgroup for all non-ray processes. The new cgroup hierarchy looks like ``` cgroup_path (e.g. /sys/fs/cgroup) | ray-node_<node_id> | | system user | | | leaf workers non-ray ``` The cgroups contain the following processes * system/leaf (all ray non-worker processes e.g. raylet, runtime_env_agent, gcs_server, ...) * user/workers (all ray worker processes) * user/non-ray (all non-ray processes migrated from cgroup_path). Note: If you're running ray inside a container, all non-ray processes running in the container will be migrated to `user/non-ray` The following controllers will be enabled * cgroup_path (cpu, memory) * ray-node_<node_id> (cpu, memory) * system (memory) The following constraints are applied * system (cpu.weight, memory.min) * user (cpu.weight) --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: elliot-barn <[email protected]>

…rm build. (ray-project#57244) For more details about the resource isolation project see ray-project#54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…ses. (ray-project#57269) This PR stacks on ray-project#57244. For more details about the resource isolation project see ray-project#54703. In the previous ray cgroup hierarchy, all processes that were in the path `--cgroup-path` were moved into the system cgroup. This changes the hierarchy to now have a separate cgroup for all non-ray processes. The new cgroup hierarchy looks like ``` cgroup_path (e.g. /sys/fs/cgroup) | ray-node_<node_id> | | system user | | | leaf workers non-ray ``` The cgroups contain the following processes * system/leaf (all ray non-worker processes e.g. raylet, runtime_env_agent, gcs_server, ...) * user/workers (all ray worker processes) * user/non-ray (all non-ray processes migrated from cgroup_path). Note: If you're running ray inside a container, all non-ray processes running in the container will be migrated to `user/non-ray` The following controllers will be enabled * cgroup_path (cpu, memory) * ray-node_<node_id> (cpu, memory) * system (memory) The following constraints are applied * system (cpu.weight, memory.min) * user (cpu.weight) --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]>

…rm build. (ray-project#57244) For more details about the resource isolation project see ray-project#54703. This PR introduces two public bazel targets from the `//src/ray/common/cgroup2` subsystem. * `CgroupManagerFactory` is a cross-platform target that exports a working CgroupManager on Linux if resource isolation is enabled. It exports a Noop implementation if running on a non-Linux platform or if resource isolation is not enabled on Linux. * `CgroupManagerInterface` is the public API of CgroupManager. It also introduces a few other changes 1. All resource isolation related configuration parsing and input validation has been moved into CgroupManagerFactory. 2. NodeManager now controls the lifecycle (and destruction) of CgroupManager. 3. SysFsCgroupDriver uses a linux header file to find the path of the mount file instead of hardcoding because different linux distributions can use different files. --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: Aydin Abiar <[email protected]>

…ses. (ray-project#57269) This PR stacks on ray-project#57244. For more details about the resource isolation project see ray-project#54703. In the previous ray cgroup hierarchy, all processes that were in the path `--cgroup-path` were moved into the system cgroup. This changes the hierarchy to now have a separate cgroup for all non-ray processes. The new cgroup hierarchy looks like ``` cgroup_path (e.g. /sys/fs/cgroup) | ray-node_<node_id> | | system user | | | leaf workers non-ray ``` The cgroups contain the following processes * system/leaf (all ray non-worker processes e.g. raylet, runtime_env_agent, gcs_server, ...) * user/workers (all ray worker processes) * user/non-ray (all non-ray processes migrated from cgroup_path). Note: If you're running ray inside a container, all non-ray processes running in the container will be migrated to `user/non-ray` The following controllers will be enabled * cgroup_path (cpu, memory) * ray-node_<node_id> (cpu, memory) * system (memory) The following constraints are applied * system (cpu.weight, memory.min) * user (cpu.weight) --------- Signed-off-by: Ibrahim Rabbani <[email protected]> Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: Aydin Abiar <[email protected]>

israbbani added 2 commits September 30, 2025 17:17

[wip] [core] (cgroups 14/n) Clean up bazel targets and expose just the

599f968

CgroupManagerFactory which constructs a cross-platform cgroup manager with selective compilation Signed-off-by: irabbani <[email protected]>

[core] Cleaning up Cgroup related bazel targets. CgroupManagerInteface

77f6b64

and CgroupManagerFactory are the only public targets. CgroupManagerFactory will delegate to the appropriate implementation for each platform. Signed-off-by: irabbani <[email protected]>

israbbani added the go add ONLY when ready to merge, run all tests label Oct 6, 2025

israbbani and others added 6 commits October 6, 2025 23:22

[core] (cgroups 14/n) Clean up bazel targets and enable cross-platform

9fd1160

build. The Cgroup subsystem only exposes CgroupManagerInterface and CgroupManagerFactory as public targets. Signed-off-by: irabbani <[email protected]>

Merge branch 'master' into irabbani/cgroups-14

a5f4b5a

Signed-off-by: irabbani <[email protected]>

Merge branch 'irabbani/cgroups-14' of github.com:ray-project/ray into…

762b5cf

… irabbani/cgroups-14 Signed-off-by: irabbani <[email protected]>

Merge branch 'master' into irabbani/cgroups-14

8d07f6f

Forgot to use clang locally

b92677e

Signed-off-by: irabbani <[email protected]>

Merge branch 'irabbani/cgroups-14' of github.com:ray-project/ray into…

a870d5a

… irabbani/cgroups-14 Signed-off-by: irabbani <[email protected]>

israbbani changed the title ~~[wip] [core] (cgroups 14/n) Clean up bazel targets and support cross-platform build.~~ [core] (cgroups 14/n) Clean up bazel targets and support cross-platform build. Oct 7, 2025

israbbani mentioned this pull request Oct 7, 2025

[Core] Resource Isolation Milestone 1 Implementation Tracker #54703

Open

16 tasks

israbbani and others added 4 commits October 7, 2025 04:32

Unused imports

e34f19b

Signed-off-by: irabbani <[email protected]>

unused includes breaking the build

55f1ec8

Signed-off-by: irabbani <[email protected]>

fixing the macos build

4c7545e

Signed-off-by: israbbani <[email protected]>

Merge branch 'master' into irabbani/cgroups-14

aef6bd8

israbbani marked this pull request as ready for review October 7, 2025 18:58

israbbani requested a review from a team as a code owner October 7, 2025 18:58

This comment was marked as outdated.

Sign in to view

ray-gardener bot added the core Issues that should be addressed in Ray Core label Oct 7, 2025

israbbani added 2 commits October 7, 2025 19:28

move operators for NoopCgroupManager

59366ce

Signed-off-by: irabbani <[email protected]>

Merge branch 'irabbani/cgroups-14' of github.com:ray-project/ray into…

44ab09e

… irabbani/cgroups-14 Signed-off-by: irabbani <[email protected]>

This comment was marked as outdated.

Sign in to view

israbbani mentioned this pull request Oct 7, 2025

[core] (cgroups 15/n) Adding a user cgroup subtree for non-ray processes. #57269

Merged

edoakes approved these changes Oct 7, 2025

View reviewed changes

israbbani and others added 2 commits October 7, 2025 15:19

Update src/ray/common/cgroup2/cgroup_manager_factory.h

0c8d8e3

Co-authored-by: Edward Oakes <[email protected]> Signed-off-by: Ibrahim Rabbani <[email protected]>

feedback

6dc39ad

Signed-off-by: irabbani <[email protected]>

up

60d77bb

Signed-off-by: israbbani <[email protected]>

Merge branch 'master' into irabbani/cgroups-14

bfd2482

edoakes enabled auto-merge (squash) October 9, 2025 19:55

github-actions bot disabled auto-merge October 9, 2025 22:54

israbbani force-pushed the irabbani/cgroups-14 branch from 8d3516c to bfd2482 Compare October 9, 2025 23:01

israbbani mentioned this pull request Oct 10, 2025

[core] Making MacOS + Java tests a little less unpleasant #57627

Closed

israbbani added 2 commits October 10, 2025 11:29

Merge branch 'master' into irabbani/cgroups-14

9174357

Merge branch 'master' into irabbani/cgroups-14

ab526bf

edoakes enabled auto-merge (squash) October 10, 2025 20:41

edoakes merged commit 7164809 into master Oct 10, 2025
6 of 7 checks passed

edoakes deleted the irabbani/cgroups-14 branch October 10, 2025 22:45

israbbani mentioned this pull request Oct 14, 2025

[Core] Remove reference counter mock for real reference counter in testing #57178

Merged

8 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[core] (cgroups 14/n) Clean up bazel targets and support cross-platform build. #57244

[core] (cgroups 14/n) Clean up bazel targets and support cross-platform build. #57244

Uh oh!

israbbani commented Oct 6, 2025 •

edited

Loading

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

edoakes left a comment

Uh oh!

edoakes Oct 7, 2025

Uh oh!

Uh oh!

edoakes Oct 7, 2025

Uh oh!

israbbani Oct 7, 2025

Uh oh!

israbbani Oct 7, 2025

Uh oh!

Uh oh!

edoakes commented Oct 7, 2025

Uh oh!

israbbani commented Oct 7, 2025 •

edited

Loading

Uh oh!

israbbani commented Oct 9, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

		// TODO(54703): Refactor the configs into a struct called CgroupManagerConfig
		// and delegate input validation and error messages to it.

[core] (cgroups 14/n) Clean up bazel targets and support cross-platform build. #57244

[core] (cgroups 14/n) Clean up bazel targets and support cross-platform build. #57244

Uh oh!

Conversation

israbbani commented Oct 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

edoakes left a comment

Choose a reason for hiding this comment

Uh oh!

edoakes Oct 7, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

edoakes Oct 7, 2025

Choose a reason for hiding this comment

Uh oh!

israbbani Oct 7, 2025

Choose a reason for hiding this comment

Uh oh!

israbbani Oct 7, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

edoakes commented Oct 7, 2025

Uh oh!

israbbani commented Oct 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

israbbani commented Oct 9, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

israbbani commented Oct 6, 2025 •

edited

Loading

israbbani commented Oct 7, 2025 •

edited

Loading