[
https://issues.apache.org/jira/browse/FALCON-1525?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14952755#comment-14952755
]
Jakub Havlík commented on FALCON-1525:
--------------------------------------
We're building an application for a big pharmaceutical company (I've already
filed a Hortonworks ticket for this, so I'm curious what will be the result).
Our goal is to have a server application which will provide REST service (doing
something on Hadoop) for UI front-end where user can create various jobs. So
when user sees that he had created a job in the front-end app we need to have a
possibility to manage that job. For example in case of an error during that job
the user can fix it on his side and restart the job via the UI. That job is of
course our private job definition as Falcon has to be hidden from user. So we
need a way to make user that when user restarts a specific job in the UI app we
are able to find a 1:1 relation with a Falcon instance and restart it. Current
Falcon solution could possibly lead to unforeseen bugs in future in some
specific corner cases. And yes I agree that frequency "minutes(1)" is weird
(but user could possibly create something like that) but I hope that from what
I written it's clear that we cannot say that we have a stable enterprise app
without this. And from my point of view if Falcon cannot say that it can behave
in a 100% predictable way with "minutes(1)" and parallel it should not have
that option.
> Add possibility to rerun process instances by an id
> ---------------------------------------------------
>
> Key: FALCON-1525
> URL: https://issues.apache.org/jira/browse/FALCON-1525
> Project: Falcon
> Issue Type: Improvement
> Components: client
> Affects Versions: 0.7
> Reporter: Jakub Havlík
> Priority: Critical
>
> As a user I need to specifically rerun individual instances by their ids.
> Doing rerun by name and time slot (where minimum resolution means minutes) is
> not usable in a corporate environment. FalconClient class needs to have a
> rerunInstance method which accepts some unique identifier for each instance.
> As it stands now there is possibility to inadvertently restart instance if it
> somehow fits into a time slot (parallel instances or instances long only a
> few seconds for example).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
