Github user d2r commented on the pull request:
https://github.com/apache/storm/pull/267#issuecomment-57811363
> I haven't enabled the acking but my understanding of the topology stats is
> aggregation of all the spouts and bolts stats.
This is not how it worked before the client-side rendering/RESTful UI
feature
was merged in.
Prior to the UI feature, showing only spout stats for acks and fails was the
way the UI operated when showing the aggregated topology stats. This
functionality was unintentionally changed at the place I linked to in the
description, and the result was that all the acks and fails including the
bolts
were summed as part of the topology stats.
This is not what our users were expecting, and I do not think it is
desirable
anyway. So this PR is to revert the functionality back to what it was
before
the client-side rendering feature in the UI.
> If I don't enable acking it doesn't show me any stats.
Yeah, the spouts need to give an ID when they emit, or else they will not
expect any ack when that ID completes processing. This ID sticks with all
downstream tuples thoughout the tuple "tree."
If you take storm.starter.clj.word_count, for example, unmodified, you will
see
exactly what you are describing, because the code has two spouts and
neither of
them give an ID when they emit a tuple. Individual bolts can still ack, but
you will not see them turn up in the spout stats.
Adding the `:id` keyword parameter fixes this.
---
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.
---
