Github user revans2 commented on the issue:
https://github.com/apache/storm/pull/2203
The changes look good an I am +1 on where they are at now.
I am a little concerned about the performance differences though. I did a
quick test with 1.x-branch (1a030939094e154a6b6d9d6b7fdf4abafd14d5b6) vs this
(8af4fcac3) on 1.x-branch.
The numbers still do not look good, and I was hoping that others might run
some tests too, just to see if they can reproduce my issues.
For ThroughputVsLatency with topology.max.spout.pending=500 I was able to
to 50,000 sentences/second without much difficulty and could just get away with
60,000. This was using the default parallelism etc.
```
./bin/storm jar storm-starter-1.2.0-SNAPSHOT.jar
org.apache.storm.starter.ThroughputVsLatency 50000 -c
topology.max.spout.pending=500
```
But with the new code it maxes out at about 30,000 sentences per second
with the same settings. I have run it multiple times to be sure it is not a
fluke.
---
