On Wed, May 16, 2018 at 4:38 PM, Sandro Bonazzola <sbonazzo@redhat.com> wrote:


2018-05-16 16:26 GMT+02:00 Yaniv Kaul <ykaul@redhat.com>:


On Wed, May 16, 2018 at 5:12 PM, Sandro Bonazzola <sbonazzo@redhat.com> wrote:


2018-05-16 15:56 GMT+02:00 Yaniv Kaul <ykaul@redhat.com>:


On Wed, May 16, 2018 at 4:04 PM, Barak Korren <bkorren@redhat.com> wrote:


On 16 May 2018 at 11:03, Sandro Bonazzola <sbonazzo@redhat.com> wrote:
Hi,
today I pushed a fix for Travis CI on oVirt Engine. While discussing the fix it has been asked why we need Travis CI in first place, isn't our Jenkins testing enough?

 
I'm wondering that as well...

What does it do that we don't do on our CI?

commit 0734e8cb9c7d1411b613f1262e81c6845b18f5d9
Author: Roman Mohr <rmohr@redhat.com>
Date:   Mon Jul 18 12:08:59 2016 +0200

    Sonarqube moved to https://sonarqube.com
    
    Change-Id: I1ce975222845bed6836464bbf73732bbcb9eff1d
    Signed-off-by: Roman Mohr <rmohr@redhat.com>

commit c703d194a4ed84541fb16b0aec73f1597ceaa05f
Author: Roman Mohr <rmohr@redhat.com>
Date:   Fri Jun 3 08:28:56 2016 +0200

    Print a keep alive message every minute on travis builds
    
    For one hour print every minute a keep alive message to stdout to keep
    the travis job alive.
    
    Sonar analysis takes more than 10 minutes. In this time period no output
    is produced. Travis assumes that builds are somehow broken if there was
    no output for more than 10 minutes.
    
    Change-Id: Ia2d048e42ee410eb72a35f2793935757308c7765
    Signed-off-by: Roman Mohr <rmohr@redhat.com>

commit fe4743ce5ffadb663d3fda58d09e9906b35b91ee
Author: Roman Mohr <rmohr@redhat.com>
Date:   Wed May 25 17:05:04 2016 +0200

    Limit sonar analysis to master branch
    
    Sonar by default does not differentiate between different branches.
    Limit the analysis to master to get untainted reports.
    
    Change-Id: Icab31f0aaf98adc88214ed70d7276f13eac188ee
    Signed-off-by: Roman Mohr <rmohr@redhat.com>

According to commit 768154debecccbc93524cd4e6c5092047d2a4eab by Roman Mohr on Sat May 7 05:39:18 2016 +0200
it's doing Sonar code analysis on travis and then upload the results to sonarqube.org
Which has been updated last time on November 21, 2017 probably due to the switch to PostgreSQL 9.5
I've no rights on that report, maybe Roman can grant some more privileges so I can try to get it back to work if needed.

 
What was broken there?

It's failing on DAO tests. 
It was failing initializing the database because it was still using postgresql 9.2 instead of 9.5.
Now it's failing on missing  uuid_generate_v1() function.

 
Can it replace some of what we do in our CI (and does it have value, as it is a mirror of published code already) ?

As far as I can tell the only gain is the sonarqube analysis of the code. I'm not sure if we can integrate this with Jenkins.

Who's looking at the analysis results?

Being it analyzing only the ovirt-engine java code I would expect backend maintainers should keep an eye on it.
Digging the mailing list history looks like this was discussed and approved in 2016
The most interested person back to that time was Piotr: are you looking at it?

No, I am not looking at it. It is something I need to improve
 




 
Y.
 



 
Y.
 




--
Barak Korren
RHV DevOps team , RHCE, RHCi
Red Hat EMEA
redhat.com | TRIED. TESTED. TRUSTED. | redhat.com/trusted

_______________________________________________
Infra mailing list -- infra@ovirt.org
To unsubscribe send an email to infra-leave@ovirt.org





--

SANDRO BONAZZOLA

ASSOCIATE MANAGER, SOFTWARE ENGINEERING, EMEA ENG VIRTUALIZATION R&D

Red Hat EMEA

sbonazzo@redhat.com   





--

SANDRO BONAZZOLA

ASSOCIATE MANAGER, SOFTWARE ENGINEERING, EMEA ENG VIRTUALIZATION R&D

Red Hat EMEA

sbonazzo@redhat.com