Github user d2r commented on the pull request:
https://github.com/apache/storm/pull/392#issuecomment-71059798
For testing this I did the following:
I modified the ExclamationTopology as follows:
- bolts report a new RuntimeException via reportError on each call to
execute
- changed topology.max.error.report.per.interval: 100 (default 10), so each
executor/thread will write up to 100 errors to ZK every 10 seconds. This is a
quick-and-dirty way of increasing write load on ZK for testing.
- Replicated the Exclamation bolt an additional ~450 times, so there are
many more bolts in the topology.
Procedure:
- Launch the topology in Inactive state, wait for it to get assigned and
stable
- go to the topology page and activate the topology (at this point the load
time is on the order of seconds)
- Refresh the topology until I see fresh errors reported for each bolt
- Wait a a minute
- Refresh the page and record load times from the Firefox network inspector.
- Repeat 4 more times
Results:
Before patch (median about a minute):
- 57s
- 1m13s
- 55s
- 51s
- 1m34s
With patch (median about 10s)
- 12s
- 6s
- 8s
- 13s
- 10s
---
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 infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
