-
Notifications
You must be signed in to change notification settings - Fork 552
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Prepare leader_balancer for node-local core assignment #16970
Conversation
This feature flag will signify nodes readiness to switch to node-local assignment of partitions to cores.
Removed piece assumes that partition.replicas contains only followers and therefore contains assignment.to, but not assignment.from (under this assumption it changes assignment.to to assignment.from). But in reality partition.replicas contains all replicas, and therefore both assignment.to and assignment.from. As other code doesn't make any assumptions about presence or absence of the leader shard from the list of replicas in the index, we can simply use the original list of replicas.
After node-local partition-core assignment becomes enabled, we can no longer rely on global replica shard info from the topic table. Use shard info from the health report instead.
/ci-repeat |
ducktape was retried in https://buildkite.com/redpanda/redpanda/builds/45876#018e1ed2-1b38-4a43-881b-208de21153b0 ducktape was retried in https://buildkite.com/redpanda/redpanda/builds/45897#018e1fab-80ab-460c-9044-3a9e27be3568 ducktape was retried in https://buildkite.com/redpanda/redpanda/builds/45897#018e1fb1-046c-4b5e-b963-528f94e5f173 |
this looks good, i am wondering what will the impact of this change be for the deployments with large number of partitions |
/cdt |
After node-local partition-core assignment becomes enabled, we can no longer rely on global replica shard info from the topic table. Use shard info from the health report instead.
Backports Required
Release Notes