-
Notifications
You must be signed in to change notification settings - Fork 3.9k
Closed
Description
Now, there is still create a new threadpool for each pipeline sync.
| executorService = Executors.newFixedThreadPool(MULTI_NODE_PIPELINE_SYNC_WORKERS); |
And this pr seems too complex and inflexible. Difficult to meet complex scenarios https://github.com//pull/4070
I have an idea for minimal changes, we can add a method that allows user to pass in the thread pool by themselves
like
public ClusterPipeline pipelined(Executor pipelineExecutor) {
return new ClusterPipeline((ClusterConnectionProvider) provider, (ClusterCommandObjects) commandObjects, pipelineExecutor);
}and make some minor modifications to other classes, like ClusterPipeline & MultiNodePipelineBase
This change is minor and can be adapted to any complex user scenario.
Any feedback is welcome. thanks
Metadata
Metadata
Assignees
Labels
No labels