[jira] [Commented] (EDGENT-397) Support failed-job-restart in IotProvider

["ASF GitHub Bot (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/EDGENT-397?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15926539#comment-15926539 ] 

ASF GitHub Bot commented on EDGENT-397:
---------------------------------------

GitHub user dlaboss opened a pull request:

    https://github.com/apache/incubator-edgent/pull/299

    [WIP] [COMMENTS?] [Edgent-112] cleanup JobMonitorApp

    As noted in the JIRA there are other issues with the class as well.
    
    - migrate submitApplication() and closeJob() to new utils.ExecutionMgmt
      - not sure about the new api:utils for this class.  Originally had it in a runtime.utils but that didn't seem right (logically it's just convenience over some api defined services), then moved it to api:execution.utils but that introduced circular dependency, so created api:utils and put it there.  Is the new api:utils OK?
    - change ApplicationServiceMXBean.submit to return the created job's id
    - add JobMonitorApp.createAndRegister (the idiom used elsewhere)
    - adjust IotProvider to use both of the above...
      - and then remove IotProvider's addition of JobMonitorProvider.  See commentary in IotProvider and EDGENT-397
    - some other general cleanup

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/dlaboss/incubator-edgent edgent112-cleanupJobMonitorApp

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/incubator-edgent/pull/299.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #299
    
----
commit c6f2ef1262336540980d51026e4b6759ff2463c7
Author: Dale LaBossiere <dl...@us.ibm.com>
Date:   2017-03-03T22:35:23Z

    [WIP] [Edgent-112] Move or remove JobMonitorApp.submitApplication
    
    As noted in the JIRA there are other issues with the class as well.
    
    - migrate submitApplication() and closeJob() to new
    runtime.utils.TopologyMgmt
    - related, change ApplicationServiceMXBean.submit to return the created
    job's id
    - add JobMonitorApp.createAndRegister (the idiom used elsewhere)
    - adjust IotProvider to use both of the above
    - remove JobMonitorApp.submit() and adjust things
    - some other general cleanup

commit ead4f712c4d300f25e2e6e1a365edcc0f46dc032
Author: Dale LaBossiere <dl...@us.ibm.com>
Date:   2017-03-04T02:15:19Z

    add gradle build parts, fix javadoc

commit 0a452bd5a63906adb879f894aa52aeaea283207c
Author: Dale LaBossiere <dl...@us.ibm.com>
Date:   2017-03-14T14:28:29Z

    runtime.utils.TopologyMgmt => execution.utils.ExecutionMgmt

commit d8f1ffbdd5ee8964d5ef1f9972839f91a5413243
Author: Dale LaBossiere <dl...@us.ibm.com>
Date:   2017-03-14T15:41:30Z

    take 2: execution.utils.ExecutionMgmt => utils.ExecutionMgmt
    
    execution.utils.ExecutionMgmt created a circular dependency created with
    topology

commit c5fe93651c1a49ce1905f6925902ed5f81c36b45
Author: Dale LaBossiere <dl...@us.ibm.com>
Date:   2017-03-15T16:10:09Z

    Remove adding JobMonitorApp to IotProvider
    
    - do early detection of JobManagerApp usage requirements

commit 905965717962bbdf79efb013f468a07537121549
Author: Dale LaBossiere <dl...@us.ibm.com>
Date:   2017-03-15T16:45:22Z

    minor cleanup

----


> Support failed-job-restart in IotProvider
> -----------------------------------------
>
>                 Key: EDGENT-397
>                 URL: https://issues.apache.org/jira/browse/EDGENT-397
>             Project: Edgent
>          Issue Type: New Feature
>          Components: Providers
>            Reporter: Dale LaBossiere
>            Assignee: Dale LaBossiere
>
> IotProvider includes the JobMonitorApp.  JobMonitorApp requires that a JobRegistryService is registered in order to receive job status events to detect failed jobs so that it can restart them.
> Neither IotProvider nor anything it uses is registering a JobRegistryService so jobs aren't restarted when using the IotProvider.
> Since IotProvider doesn't document that it adds JobMonitorApp and it doesn't state that it provides failed app restart capability, I've decided this is really a feature request.  The original non-restarting behavior must be preserved as the default behavior.
> Might want to get a better understanding of failure handling before exposing job-restart via IotProvider - see EDGENT-382.
> See also EDGENT-396 and EDGENT-112
> Note, JobMonitorAppTest verifies the app in isolation however there isn't an analogous test in an IotProvider context. IotProviderTest should be enhanced accordingly.  With a couple of tweaks it can leverage some of JobMonitorAppTest.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)