Github user zhuoliu commented on a diff in the pull request:
https://github.com/apache/storm/pull/765#discussion_r40803217
--- Diff: storm-core/src/clj/backtype/storm/daemon/worker.clj ---
@@ -123,12 +123,13 @@
port (:port worker)
storm-cluster-state (:storm-cluster-state worker)
prev-backpressure-flag @(:backpressure worker)]
- (if executors
- (if (reduce #(or %1 %2) (map #(.get-backpressure-flag %1)
executors))
- (reset! (:backpressure worker) true) ;; at least one
executor has set backpressure
- (reset! (:backpressure worker) false))) ;; no executor has
backpressure set
+ (when executors
+ (reset! (:backpressure worker)
+ (or @(:transfer-backpressure worker)
--- End diff --
Previously, we have lowWaterMark function as "do nothing" for worker
transfer queue's callback, which will not turn off the backpressure when the
worker's queue population is below the lowWaterMark. On the other hand, it is
possible that all executors all have population above above lowWaterMark but
the worker's queue population is still above highWaterMark, and the
backpressure flag gets cleared "too often". Anyway, I think this is a good fix,
together with the _throttleOn flag in DisruptorQueue.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
