Assigned
Status Update
No update yet.
Comments
ba...@google.com
ka...@google.com
ka...@google.com
I had the same problem!
I could solve it by putting the .aidl file in an aidl directory.
Look at the attached screenshot to see the project structure.
I could solve it by putting the .aidl file in an aidl directory.
Look at the attached screenshot to see the project structure.
Description
We have container workloads running under GKE that require the ability for a parent process to continue running in the event that a subprocess it launches is OOMKilled.
This has been working fine under CGROUPSv1 where the default behavior is for only the OOMing process to be killed. However, with CGROUPSv2, which we're being forced to adopt on GKE within the next few versions, the default behavior is for ALL processes in the container to be killed when one of them OOMs.
This new config flag, added in upstream Kubernetes v1.32 allows the ability to control whether CGROUPSv2 kills just the OOMing process, or all processes in the container.
Not having this control once CGROUPv1 nodes are no longer available may make GKE unusable for some workloads until significant rework has been done.