HSEARCH-3084 Initialize and close index managers / backends in parallel #2109
Commits on Oct 2, 2019
-
HSEARCH-3084 Retrieve Elasticsearch metadata (mapping + settings) usi…
…ng a single REST API call
-
HSEARCH-3084 Make failure collectors thread-safe
Necessary because of ES schema validation in particular.
-
HSEARCH-3084 Delay shard starting until after all shards are created …
…in Lucene index managers
-
HSEARCH-3084 Delay directory creation/initialization until shards are…
… started in Lucene index managers
-
HSEARCH-3084 Do not display the "Context: " prefix when rendering Eve…
…ntContexts by default
-
HSEARCH-3084 Handle index names more consistently in Lucene index man…
…agers Make sure to always include the shard ID, in particular.
-
HSEARCH-3084 Simplify Lucene write orchestrator creation
There's no reason to go through an interface exposed in a separate package: the consumer and the implementation of the createWriteOrchestrator() method are located in the same package.
-
-
HSEARCH-3084 Simplify BatchingExecutor
1. Get rid of the Phaser and use much simpler code. 2. Expose a CompletableFuture<?> to wait for full completion (will be useful in the next commits).
-
HSEARCH-3084 Pre-stop backends and index managers in parallel
Mainly, this means that on shutdown, we'll stop accepting works for all backends/index managers immediately, and *then* we'll wait for ongoing works to complete. The improvement over the previous behavior will probably be negligible, but what's more important is that the path towards an API for explicitly starting/pre-stopping/stopping index managers is now clearer.
-
HSEARCH-3084 Add basic unit tests for BatchingExecutor
It's not a lot, but it's better than nothing.
-
HSEARCH-3084 Always execute all worksets in BatchingExecutor, even if…
… a previous workset failed The problem is unrelated to this PR, but was detected thanks to the tests I added to make sure the new implementation works correctly.
