Scheduling and tracking control plane operations for distributed storage systems

摘要

A system that implements distributed storage may schedule and track control plane operations for performance at the distributed storage service. Information may be maintained for control plane events detected at a distributed storage system. Resource utilization for currently performing control plane operations and currently scheduled control plane operations of the distributed storage system may be determined. The information about detected control plane events may be analyzed to schedule control plane operations to be performed in response to detecting the control plane events. As part of scheduling control plane operations, resource constraints may be applied to the determine resource utilization for the distributed storage system.

主权项

1. A system, comprising:
a plurality of compute nodes implementing a distributed storage service, wherein the distributed storage service comprises:
a plurality of task handlers, configured to perform a plurality of operations for the distributed storage service;a control plane event registry, configured to:
receive a plurality of control plane event registrations from one or more of the plurality of compute nodes, wherein each control plane event registration indicates the detection of a control plane event at the distributed storage service, wherein each detected control plane event corresponds to one or more of the plurality of control plane operations which are to be performed in response to detecting the control plane event; andstore information about the detected control plane events indicated by the plurality of control plane event registrations; anda control plane task scheduler, configured to:
determine current resource utilization for currently performing control plane operations and currently scheduled control plane operations of the distributed storage service;analyze the information about the detected control plane events to schedule at least one of the one or more control plane operations corresponding to the detected control plane events to be performed, based, at least in part, on applying one or more resource constraints for performing the one or more control plane operations to the determined current resource utilization; anddirect at least one task handler of the plurality of task handlers to perform the scheduled at least one control operation.