inline From: [email protected] [mailto:[email protected]] On Behalf Of Lori Jakab Sent: Monday, July 11, 2016 4:54 PM To: Mainzer, Gal <[email protected]> Cc: Jamo Luhrsen <[email protected]>; discuss opendaylight <[email protected]>; Lori Jakab <[email protected]>; [email protected]; [email protected]; odl lispflowmapping dev <[email protected]> Subject: Re: [OpenDaylight Discuss] [OpenDaylight Infrastructure] External CI
Hi Gal, On Sun, Jul 10, 2016 at 12:12 AM, Mainzer, Gal <[email protected]<mailto:[email protected]>> wrote: This is really cool. We are indeed trying to enhance the ODL CI with our lab resources as well. That's great, keep us posted on this thread, since this is new for everyone. Lori, can you share a bit what you did? Is your CI based on jenkins? Right now it is just a simple python script using the pygerrit library. We haven't decided yet if we'll go with Jenkins, or simple homegrown scripts. At least in the beginning we'll do just some scripting, since Jenkins seems overkill, but we will likely use the Robot framework for tests, so that we can leverage some of the work we already did in the integration/test repo. [GM] we faced the same thoughts. At the end, we decided to go with existing ODL infrastructure and flows – this will definitely give us the ability to contribute back important enhancements and improve the CI process. Also, in such cases where several different groups wishes to run external CI and contribute the results, we will be able to share the effort and have multiple external CIs run different sets of tests. However, as Andy mentioned previously on this thread, we could use Jenkins with the Gerrit Trigger plugin to interface with ODL Gerrit, instead of our python script. Are you planning to publish logs/reports in case of failure so that patch writer will be able to analyze the failure? We will decide this based on the tests themselves, once we have them. The test infra is behind firewall, so there won't be a web page with extra info on the failure. As a result, at least some indication of what went wrong should be in the failure comment. We will experiment with different approaches. [GM] our thoughts on this is that it may become problematic if we put the external CI as gating for patches, therefore we will need to publish the result somehow to allow others to debug their failures. We should work together on this and try to create an external web server that will hold the logs – the best part is that existing JJB scripts contain web server result publication so it shouldn’t be that hard to make it work locally. HTH, -Lori -----Original Message----- From: [email protected]<mailto:[email protected]> [mailto:[email protected]<mailto:[email protected]>] On Behalf Of Jamo Luhrsen Sent: Friday, July 08, 2016 10:04 PM To: Lori Jakab <[email protected]<mailto:lorand.jakab%[email protected]>> Cc: discuss opendaylight <[email protected]<mailto:[email protected]>>; Lori Jakab <[email protected]<mailto:[email protected]>>; [email protected]<mailto:[email protected]>; [email protected]<mailto:[email protected]>; odl lispflowmapping dev <[email protected]<mailto:[email protected]>> Subject: Re: [OpenDaylight Discuss] [OpenDaylight Infrastructure] External CI On 07/08/2016 11:19 AM, Lori Jakab wrote: > Hi, > > On Fri, Jun 10, 2016 at 9:06 AM, Lori Jakab > <[email protected]<mailto:lorand.jakab%[email protected]> > <mailto:[email protected]<mailto:lorand.jakab%[email protected]>>> > wrote: > > On Wed, Jun 8, 2016 at 6:44 PM, Jamo Luhrsen > <[email protected]<mailto:[email protected]> > <mailto:[email protected]<mailto:[email protected]>>> wrote: > > Let's bring this topic up in the Integration call tomorrow as well. > > Lori, Andy usually attends that call if you had the time to make it > and wanted to have any discussion. > > > Hi JamO, > > Sorry I missed that call. I will look into this when I get a chance with > the info from Andy, and follow up on this thread > if I make progress or hit issues. > > > This work has been (and is still) on the back-burner, but I just > wanted to report a small progress. Based on the example script from > the pygerrit repo [0] I have a script that is listening for gerrit > stream events, filters for lispflowmapping new patch set events, and adds a > (non-voting for now comment). Here's an example [1], look at the second > comment :D it was added by the "external CI". > > From here, we'll work on defining some basic tests, and adding voting. One > step at a time. This is awesome progress, Lori. Thanks for sharing the status and script. I think HPE is also considering this. JamO > -Lori > > [0] https://github.com/sonyxperiadev/pygerrit/blob/master/example.py > [1] https://git.opendaylight.org/gerrit/#/c/41572/ > > > Thanks for all the input. > > Regards, > -Lori >
_______________________________________________ Discuss mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/discuss
