On Wed, Jun 22, 2016 at 12:28 PM, Eyal Edri <eedri@redhat.com> wrote:
After the recent merge of testing install of ovirt-cockpit-dashboard to 3.6,
the 4.0 tests failed also. [1]

They should not have failed. The only reason they could have failed is if there were added deps - and there were - to the dashboard.
Specifically, hosted-engine-setup was added as a dep, bringing with it a huge amount of other packages (virt-viewer, which although I requested was not removed, which brings spice, which brings GTK...) - overall, ~500 RPMs (!).

This is the reason I thought of abandoning this patch - which I think I've commented on in the patch itself.


And then I found out that some of the 4.0 tests are linked to 3.6 still, is this intentional?

Yes, for two reasons:
1. It allows less code duplication.
2. It allows us to test 4.0 with v3 API.
3. It allows us to compare 4.0 to 3.6.x.
 
Should we now create a new separate 4.0 test or change the link to master?

We need at some point to add v4 API tests to 4.0.
Y.
 

lrwxrwxrwx. 1 eedri eedri 64 Jun 22 00:12 001_initialize_engine.py -> ../../basic_suite_master/test-scenarios/001_initialize_engine.py
lrwxrwxrwx. 1 eedri eedri 53 Jun 22 00:12 002_bootstrap.py -> ../../basic_suite_3.6/test-scenarios/002_bootstrap.py
lrwxrwxrwx. 1 eedri eedri 56 Jun 22 00:12 004_basic_sanity.py -> ../../basic_suite_3.6/test-scenarios/004_basic_sanity.py



[1] http://jenkins.ovirt.org/job/ovirt_4.0_system-tests/72/console

--
Eyal Edri
Associate Manager
RHEV DevOps
EMEA ENG Virtualization R&D
Red Hat Israel

phone: +972-9-7692018
irc: eedri (on #tlv #rhev-dev #rhev-integ)