[openstack-dev] [QA][Neutron][3rd Party Testing] Methodology for 3rd party

trinath.somanchi at freescale.com trinath.somanchi at freescale.com
Wed Feb 5 06:36:29 UTC 2014


Hi -

I'm a new bee here..

Can anyone guide me on setting up a new 3rd party Testing account.

How it is useful ?
What the machine requirements?
How testing is automated ?
How to post back +/- 1 to Jenkins ?
What packages are to be installed ?

Kindly help me understand them.

Thanks in advance
--
Trinath Somanchi - B39208
trinath.somanchi at freescale.com | extn: 4048

From: Miguel Angel [mailto:miguelangel at ajo.es]
Sent: Wednesday, February 05, 2014 11:59 AM
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [QA][Neutron][3rd Party Testing] Methodology for 3rd party

Interesting points here, I agree with Akihiro, some components
leave services, and left over settings over the system even when
shut down (I know of neutron net namespaces, .. etc..).

+1 to Akihiro proposals for a fresh-vm.


---
irc: ajo / mangelajo
Miguel Angel Ajo Pelayo
+34 636 52 25 69
skype: ajoajoajo

2014-02-05 Akihiro Motoki <motoki at da.jp.nec.com<mailto:motoki at da.jp.nec.com>>:
Hi,

I think it is better to use a fresh VM to run tests.
When running tempest scenario tests, there is a case
where some resources can not be cleanup properly.
It happens when some test fails of course.

I think 10 minutes is not too long.
It requires more than 30 minutes until gate jobs
on openstack-ci report test results.
10 minutes is fast enough compared to this time.

Other ways to speed up the testing are:
- to instsall dependecy packages in advance
- to create PyPI mirror
- to clone required git repos in advance and just sync when testing
 From my experience the first one will contribute most to save time.

Thanks,
Akihiro

(2014/02/05 10:24), Franck Yelles wrote:
> Hello,
>
> I was wondering how everyone was doing 3rd party testing at the moment
> when it comes to the process.
> It takes me around 10 minutes for me to do a +1 or -1.
>
> my flow is the following:
> (I only use Jenkins for listening to the "feed")
> 1) a job is triggered from Jenkins.
> 2) a VM is booted
> 3) the devstack repo is clone
> 4) the patch is applied
> 5) stack.sh is run (longest time is here)
> 6) the test are run
> 7) the result is posted
> 8) the VM is destroyed
>
> I am looking for ways to speed up the process.
> I was thinking of keeping the stack.sh up;  and follow this
>
> 1) Shutdown the affected component  (neutron, etc..)
> 2) apply the patch
> 3) restart the component
> 4) run the test
> 5) post the result
> 6) shutdown the affected component
> 7) remove the patch
> 8) restart the component
>
> What are you thoughts ?
> Ideally I would like to achieve a sub 3 minutes.
>
> Thanks,
> Franck
>
> _______________________________________________
> OpenStack-dev mailing list
> OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140205/5ce24bc8/attachment.html>


More information about the OpenStack-dev mailing list