Do not treat replica as unassigned if primary recently created and unassigned time is below a threshold. #112066
+271 −80
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
elasticsearchmachine added v8.16.0 Team:Data Management 
| Pinging @elastic/es-data-management (Team:Data Management) |
| Hi @parkertimmins, I've created a changelog YAML for you. |
dakrone requested changes Aug 21, 2024
...lasticsearch/cluster/routing/allocation/shards/ShardsAvailabilityHealthIndicatorService.java Outdated Show resolved Hide resolved
...lasticsearch/cluster/routing/allocation/shards/ShardsAvailabilityHealthIndicatorService.java Outdated Show resolved Hide resolved
...lasticsearch/cluster/routing/allocation/shards/ShardsAvailabilityHealthIndicatorService.java Outdated Show resolved Hide resolved
- pass now millis instead of clock - update setting with addSettingsUpdateConsumer - simplify cutoff condition logic
- need to add settings to mocked ClusterService
parkertimmins commented Aug 22, 2024
...lasticsearch/cluster/routing/allocation/shards/ShardsAvailabilityHealthIndicatorService.java Outdated Show resolved Hide resolved
parkertimmins commented Aug 22, 2024
...lasticsearch/cluster/routing/allocation/shards/ShardsAvailabilityHealthIndicatorService.java Outdated Show resolved Hide resolved
dakrone approved these changes Aug 26, 2024
...lasticsearch/cluster/routing/allocation/shards/ShardsAvailabilityHealthIndicatorService.java Outdated Show resolved Hide resolved
cbuescher pushed a commit to cbuescher/elasticsearch that referenced this pull request Sep 4, 2024
…lastic#112066) Changes the way we calculate if all replicas are unassigned when primary is recently created. This change will only be used in serverless, not in stateful. When a primary is new, if the primary is active, but the replica is unassigned for less than a buffer time period, do not treat is as unassigned. Control time period through health.shards_availability.replica_unassigned_buffer_time setting.
parkertimmins added a commit that referenced this pull request Sep 12, 2024
Increase the default value of health.shards_availability.replica_unassigned_buffer_time to 5 seconds. This values in the identification of unavailable shards on serverless. Increasing the value to 5s keep more shards from going red transiently, while still being low enough to go red quickly if there is an actual availability issue. Related to #112066
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit. This suggestion is invalid because no changes were made to the code. Suggestions cannot be applied while the pull request is closed. Suggestions cannot be applied while viewing a subset of changes. Only one suggestion per line can be applied in a batch. Add this suggestion to a batch that can be applied as a single commit. Applying suggestions on deleted lines is not supported. You must change the existing code in this line in order to create a valid suggestion. Outdated suggestions cannot be applied. This suggestion has been applied or marked resolved. Suggestions cannot be applied from pending reviews. Suggestions cannot be applied on multi-line comments. Suggestions cannot be applied while the pull request is queued to merge. Suggestion cannot be applied right now. Please check back later.
Changes the way we calculate if all replicas are unassigned when primary is recently created.
Related to #107794 which did not treat replica as unassigned if primary was not yet active.
This changes extend the time when replicas are not treated as unassigned to a buffer time period after the primary has become active. This buffer time period is controlled through the setting
health.shards_availability.replica_unassigned_buffer_time. This is only used in serverless; on stateful the behavior remains the same where new unassigned replicas will only not be treated was unassigned while the primary is not yet active.