Scheduler "stops" after a while

[markretallack]

Environment details:

• AWS EKS 1.29
• Tag 7863e27 for scheduler, controller, invoker

Steps to reproduce the issue:

1. Deploy using scheduler etc...
2. deploy some functions (cron based etc...

Provide the expected results and outputs:

The system works normally

Provide the actual results and outputs:

After a while ( about an hour). The system becomes unstable.

In the controller log:

[2024-05-21T13:00:09.119Z] [ERROR] [#tid_kDcLqa1uLqzR7GhLLdejqhGeS3mbeHQo] [] Failed to recreate queue for dataspace/ncarpark/carpark@0.0.1, no scheduler endpoint available

Also seeing this in the controller log:

[2024-05-21T13:00:06.174Z] [WARN] [#tid_kDcLqa1uLqzR7GhLLdejqhGeS3mbeHQo] [] The whisk/queue/dataspace/dataspace/carpark/carpark/leader is deleted from ETCD, but there are still unhandled activations for this action, try to create a new queue

In the scheduler log I am seeing this:

[2024-05-21T13:00:09.876Z] [WARN] [#tid_sid_unknown] [EtcdWorker] a lease is expired while registering an initial data whisk/queue/dataspace/dataspace/carpark/carpark/leader, reissue it: io.grpc.StatusRuntimeException: NOT_FOUND: etcdserver: requested lease not found

And also:

[2024-05-21T13:00:10.195Z] [WARN] [#tid_sid_unknown] [EtcdWorker] a lease is expired while registering an initial data whisk/scheduler/0, reissue it: io.grpc.StatusRuntimeException: NOT_FOUND: etcdserver: requested lease not found

Not sure where to look for this issue

[markretallack]

My current solution is to disable the new scheduler until I can find the issue.