After the recent merge of testing install of ovirt-cockpit-dashboard to 3.6, the 4.0 tests failed also. [1]

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

Should we now create a new separate 4.0 test or change the link to master?

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)

On Wed, Jun 22, 2016 at 2:27 PM, Yaniv Kaul <ykaul(a)redhat.com&gt; wrote: > > > On Wed, Jun 22, 2016 at 12:28 PM, Eyal Edri <eedri(a)redhat.com&gt; 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 > (!). > They failed because 4.0 was linked to 3.6, I've posted this fix: https://gerrit.ovirt.org/#/c/59603/ When we'll make it work for 4.0, we'll restore the link.

> > 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. > I'm fully aware of these, but the fact we might have different tests / deps for different version will require us at some point to split it. And later on do some refactor to make sure we're sharing what we can to all tests. > > >> 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) >> > > -- 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)

On Wed, Jun 22, 2016 at 2:55 PM, Yaniv Kaul <ykaul(a)redhat.com&gt; wrote: > > > On Wed, Jun 22, 2016 at 2:49 PM, Eyal Edri <eedri(a)redhat.com&gt; wrote: > >> >> >> On Wed, Jun 22, 2016 at 2:27 PM, Yaniv Kaul <ykaul(a)redhat.com&gt; wrote: >> >>> >>> >>> On Wed, Jun 22, 2016 at 12:28 PM, Eyal Edri <eedri(a)redhat.com&gt; 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 (!). >>> >> >> >> They failed because 4.0 was linked to 3.6, I've posted this fix: >> https://gerrit.ovirt.org/#/c/59603/ >> When we'll make it work for 4.0, we'll restore the link. >> > > So 4.0 was broken? Because I've tested on BOTH 3.6 and master - see my > comment @ https://gerrit.ovirt.org/#/c/58775/ > Yes, was missing packages: http://jenkins.ovirt.org/job/ovirt_4.0_system-tests/72/testReport/junit/(... --> Processing Dependency: virt-viewer for package: ovirt-hosted-engine-setup-2.0.0.3-0.0.master.20160616133444.gitd50de9a.el7.centos.noarch --> Processing Dependency: socat for package: ovirt-hosted-engine-setup-2.0.0.3-0.0.master.20160616133444.gitd50de9a.el7.centos.noarch --> Processing Dependency: python-libguestfs for package: ovirt-hosted-engine-setup-2.0.0.3-0.0.master.20160616133444.gitd50de9a.el7.centos.noarch --> Finished Dependency Resolution You could try using --skip-broken to work around the problem You could try running: rpm -Va --nofiles --nodigest lago.ssh: DEBUG: Command a953e492 on lago_basic_suite_4_0_host0 errors: Error: Package: ovirt-hosted-engine-setup-2.0.0.3-0.0.master.20160616133444.gitd50de9a.el7.centos.noarch (alocalsync) Requires: virt-viewer Error: Package: cockpit-ovirt-dashboard-0.10.5-1.0.0.el7.centos.noarch (alocalsync) Requires: cockpit Error: Package: ovirt-hosted-engine-setup-2.0.0.3-0.0.master.20160616133444.gitd50de9a.el7.centos.noarch (alocalsync) Requires: socat Error: Package: ovirt-hosted-engine-setup-2.0.0.3-0.0.master.20160616133444.gitd50de9a.el7.centos.noarch (alocalsync) Requires: python-libguestfs lago.utils: ERROR: Error while running thread Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/lago/utils.py", line 53, in _ret_via_queue queue.put({'return': func()}) File "/home/jenkins/workspace/ovirt_4.0_system-tests/ovirt-system-tests/basic_suite_4.0/test-scenarios/002_bootstrap.py", line 163, in _install_cockpit_ovirt_on_host nt.assert_equals(ret.code, 0, '_install_cockpit_ovirt_on_host(): failed to install cockpit-ovirt-dashboard on host %s' % host) File "/usr/lib64/python2.7/unittest/case.py", line 551, in assertEqual assertion_func(first, second, msg=msg) File "/usr/lib64/python2.7/unittest/case.py", line 544, in _baseAssertEqual > Y. > > >> >>> >>> 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. >>> >> >> I'm fully aware of these, but the fact we might have different tests / >> deps for different version will require us at some point to split it. >> And later on do some refactor to make sure we're sharing what we can to >> all tests. >> >> >>> >>> >>>> 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) >>>> >>> >>> >> >> >> -- >> 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) >> > > -- 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)