[richfaces-planning-issues] [JBoss JIRA] (RFPL-2344) Implement POC of deploying only once to container
Juraj Húska (JIRA)
issues at jboss.org
Wed Jul 9 11:13:24 EDT 2014
[ https://issues.jboss.org/browse/RFPL-2344?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12983570#comment-12983570 ]
Juraj Húska edited comment on RFPL-2344 at 7/9/14 11:11 AM:
------------------------------------------------------------
[Arquillian-suite-extension|https://github.com/ingwarsw/arquillian-suite-extension], is doing exactly what we want here. One deployment per test suite.
I have incorporated the test suite in [this|https://github.com/richfaces/richfaces-qa/commit/cc0f412a38ae2624d002e2ce238ec6987b3748b2] branch and commit.
Tested it on these jobs:
* [WildFly 8.1|https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/RichFaces/view/RF-QE/job/richfaces-4.5-metamer-one-deployment-qe/]
* [EAP 6|https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/RichFaces/view/RF-QE/job/richfaces-wfk-2.6-one-deployment-qe/]
It works as expected, and brings these advantages:
* decrease of tests execution
** it is not dramatic, but at least something, e.g. job for WildFly took *half an hour* less, job for EAP took instead of 56 min only 40 min. (those are different test suites)
* more clear test log
** no mess from container restart and redeploy of Metamer
* memory consumption improved
** I guess we can test with less heap and PermGen space
My question [~ppitonak], [~jstefek], [~manovotn], [~mtomasek] is whether it is worth incorporating it ?
IMHO yes, the only thing which is bottleneck is adding third party dependency (Its simple extension, quite popular in community, well maintained).
Thanks for feedback.
was (Author: jhuska):
[Arquillian-suite-extension|https://github.com/ingwarsw/arquillian-suite-extension], is doing exactly what we want here. One deployment per test suite.
I have incorporated the test suite in [this|https://github.com/richfaces/richfaces-qa/commit/cc0f412a38ae2624d002e2ce238ec6987b3748b2] branch and commit.
Tested it on these jobs:
* [WildFly 8.1|https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/RichFaces/view/RF-QE/job/richfaces-4.5-metamer-one-deployment-qe/]
* [EAP 6|https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/RichFaces/view/RF-QE/job/richfaces-wfk-2.6-one-deployment-qe/]
It works as expected, and brings these advantages:
* decrease of tests execution
** it is not dramatic, but at least something, e.g. job for WildFly took *half an hour* less, job for EAP took instead of 56 min only 40 min. (those are different test suites)
* more clear test log
** no mess from container restart and redeploy of Metamer
* memory consumption improved
** I guess we can test with less heap and PermGen space
My question [~ppitonak], [~jstefek], [~manovotn], [~mtomasek] is whether it is worth incorporating it ?
IMHO yes, the only thing which is bottleneck is adding third party dependency (Its simple extension, quite popular in community).
Thanks for feedback.
> Implement POC of deploying only once to container
> -------------------------------------------------
>
> Key: RFPL-2344
> URL: https://issues.jboss.org/browse/RFPL-2344
> Project: RichFaces Planning
> Issue Type: Task
> Security Level: Public(Everyone can see)
> Components: QE
> Affects Versions: 4.3.0.M1
> Reporter: Juraj Húska
> Assignee: Juraj Húska
>
> Try to implement POC of this approach:
> # unzip container
> # copy application war to the directory from where it is automatically deployed
> # comment deployment section of tests
> # do that at first for showcase
> # run test suite with and without the trick and see the results
--
This message was sent by Atlassian JIRA
(v6.2.6#6264)
More information about the richfaces-planning-issues
mailing list