thinkharderdev commented on code in PR #184:
URL: https://github.com/apache/arrow-ballista/pull/184#discussion_r962305623
##########
ballista/rust/scheduler/src/state/executor_manager.rs:
##########
@@ -383,15 +430,42 @@ impl ExecutorManager {
Ok(())
}
- /// Initialize the set of executor heartbeats from storage
- pub(crate) async fn init_executor_heartbeats(&self) -> Result<()> {
+ pub(crate) async fn save_dead_executor_heartbeat(
Review Comment:
I'm a little confused as to why we keep the dead executors around in the
state.
##########
ballista/rust/scheduler/src/state/execution_graph.rs:
##########
@@ -485,8 +490,139 @@ impl ExecutionGraph {
self.output_locations.clone()
}
+ /// Reset running and completed stages on a given executor
Review Comment:
There is a lot of subtle logic here. I went through it the best I could and
it seems ok but we should really have some test coverage for this.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
