You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@storm.apache.org by "Rick Kellogg (JIRA)" <ji...@apache.org> on 2015/10/09 02:20:27 UTC

[jira] [Updated] (STORM-116) [UI] Sort bolt/spout in order that wrote in toplogy

     [ https://issues.apache.org/jira/browse/STORM-116?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Rick Kellogg updated STORM-116:
-------------------------------
    Component/s: storm-core

> [UI] Sort bolt/spout in order that wrote in toplogy
> ---------------------------------------------------
>
>                 Key: STORM-116
>                 URL: https://issues.apache.org/jira/browse/STORM-116
>             Project: Apache Storm
>          Issue Type: New Feature
>          Components: storm-core
>            Reporter: James Xu
>
> https://github.com/nathanmarz/storm/issues/129
> I think bolt/spout in storm UI should be sorted in the order that I wrote in topology.
> For example, I wrote
> builder.setBolt("SplitSentence", new SplitSentenceBolt(), 3)
>         .shuffleGrouping("SentenceSpout");
> builder.setBolt("Exclamation", new ExclamationBolt(), 2)
>         .shuffleGrouping("SplitSentence");
> In UI, they are sorted by their ID like this:
> Exclamation    2 .....
> SplitSentence 3 .....
> To fix this issue, I think it is good to add left column that shows its raw number like this:
> # / id / parallelism ....
> 1 / splitsentence / 3 ...
> 2 / exclamation / 2 ....
> ----------
> xumingming: things will get complicated because the topology could be a graph. maybe it could be implemented in another project to have a more 'visualized' view of topology. e.g. the topology is shown as a connected node of bolt, and you can click on each node to check the details of each bolt, task etc.
> ----------
> mrflip: Barring DRPC links, is it the case that a topology is always a DAG? If so, a "topological sort" is always possible (with a tiebreaker to make it unique), and a depth-first (1, 1.a, 1.a.i, 1.b, 2) or breadth-first (1, 2, 1.a, 1.b, 1.a.1) enumeration of the graph would give stable sortable labels.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)