[GitHub] storm issue #2241: STORM-2306 : Messaging subsystem redesign.

[roshannaik <gi...@git.apache.org>]

Github user roshannaik commented on the issue:

    https://github.com/apache/storm/pull/2241
  
    @revans2 
    - Perhaps best to wait till I introduce the sleep strategy before  retry-ing those lower throughput runs. This is my top priority right now.
    - Great inputs on the max.spout.pending. My thoughts on the same:
    
    1. Yes. We should not remove it until netty issue is fixed. Good to aim for fixing this in Storm 2.0 as max.spout.pending is only useful in ACK mode.
    2.  Not clear about this issue
    3. Yes the LoadAware model makes sense from a logical standpoint. Its perf bothered me enough that and I felt it might be hurting as well as helping the larger cause. I anticipated some "backpressure" on this decision. I am ok to re-enable it and document this as something to consider for perf tuning. For some of this perf testing, IMO we need  be conscious when to enable/disable it. Would be great if someone can look into improving its perf.
    4. Will address the timers issue in a separate post.
    5. Unclear about some aspects of this issue, but deprecation of cyclic topos seems reasonable to me.


---
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 infrastructure@apache.org or file a JIRA ticket
with INFRA.
---