[Starlingx-discuss] Request freeze on commits

[Miller, Frank]

A new build was started 30 minutes ago by Davlet to pick up a fix for the docker image build [1].  Nicolae would it be possible to schedule another sanity today when this build is available?

Once the sanity results from earlier today are available, we need to determine if the root cause is the same root cause as before [2].  If not we will need additional fixes/plan to address today's sanity error.

In parallel, Davlet is working on getting the monolithic build to generate helm charts and docker images so we have a backup plan if issues continue to be seen that can be mapped back to build layering.

Frank
[1] https://review.opendev.org/#/c/727218/
[2] https://bugs.launchpad.net/starlingx/+bug/1877548


-----Original Message-----
From: Saul Wold <sgw at linux.intel.com> 
Sent: Tuesday, May 12, 2020 11:21 AM
To: Miller, Frank <Frank.Miller at windriver.com>; starlingx-discuss <starlingx-discuss at lists.starlingx.io>
Subject: Re: [Starlingx-discuss] Request freeze on commits

+100!!!!

I was about to write this same email, Sanity testing is still failing from the last night's load with the Layered build (email will come after Test meeting).

So the fix that was applied yesterday for the helm chart broke the build?  Or was there some other commit?  I thought we had already declared a full stop last week except for required fixes to address the Sanity failures only.

Are you seeing the same failure with the helm charts as the Intel test team?

Is there a way to get the Monolithic helm charts build independent of the ISO? I think this is what Austin was getting at.

Sau!


On 5/12/20 8:06 AM, Miller, Frank wrote:
> StarlingX Community:
> 
> We have struggled for the past 2+ weeks to get a sane load.  At this 
> point I have no choice but to ask all Cores to freeze commits and not 
> allow any commits to merge unless they address a build or sanity issue.
> 
> Last night's docker image build failed due to a recent commit and the 
> developer is providing a fix.  But we also have issues prior to this 
> in sanity due to the build layering method pulling in the wrong 
> version of a package.  If the current fixes do not result in sanity 
> passing then I'll ask the build team to revert to the monolithic build 
> until a better solution can be found.
> 
> **
> 
> Frank
> 
> Project Lead for Build
> 
> 
> _______________________________________________
> Starlingx-discuss mailing list
> Starlingx-discuss at lists.starlingx.io
> http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
>