On Mon, May 30, 2016 at 9:35 AM, Eyal Edri <eedri(a)redhat.com> wrote:
both check-patch and check-merge are using the following rpmbuild
params:
rpmbuild \
-D "_rpmdir $PWD/output" \
-D "_topmdir $PWD/rpmbuild" \
-D "release_suffix ${SUFFIX}" \
-D "ovirt_build_ut $BUILD_UT" \
-D "ovirt_build_extra_flags $EXTRA_BUILD_FLAGS" \
-D "ovirt_build_quick 1" \
--rebuild output/*.src.rpm
I think the -D "ovirt_build_quick 1" skips GWT compilation or at
least some of the permutations, so if the errors was related to GWT it
explains the problem.
I suggest to change check-merge to include the basic GWT compilation with a
single permutation so we'll catch these at least after merge (if
its
doesn't take too long, worth considering adding to
check-patch.sh).
Yes, ovirt_build_quick skips GWT compilation completely, we should use
ovirt_build_minimal which compiles GWT for English and Firefox only.
This requires a patch to ovirt-engine automation/check-merge.sh file on
master, ovirt-engine-4.0 and ovirt-engine-3.6.
Here's the patch for master:
https://gerrit.ovirt.org/58246
On Mon, May 30, 2016 at 10:25 AM, Martin Perina <mperina(a)redhat.com>
wrote:
>
>
> On Mon, May 30, 2016 at 9:05 AM, Barak Korren <bkorren(a)redhat.com> wrote:
>
>> On 30 May 2016 at 09:51, Martin Perina <mperina(a)redhat.com> wrote:
>> > Hi,
>> >
>> > I just verified that local build is is broken by the same error in
>> > HostedEngineDeployConfiguration.
>> > I will post a fix for that, but upgrade-job build is correct (raised
>> build
>> > failure) and check-merged jobs
>> > are broken (producing success although build is broken), could you
>> please
>> > fix them asap?.
>> >
>>
>> Since those are std-CI jobs, I've no clue what they do and how to fix
>> them, and also no permissions to do so (I'm not an engine maintainer).
>> Everything they do is in the automation dir.
>>
>> You can try to run them locally as well, see [1]
>>
>> [1]:
>>
http://ovirt-infra-docs.readthedocs.io/en/latest/CI/Build_and_test_standa...
>>
>> It is an interesting problem though, what could make the build succeed
>> inside mock but fail otherwise?
>>
>
> There's no way how those builds could be successful regardless of using
> mock, GWT build had always to fail without patch [2], so the issue is
> somewhere inside CI/Jenkins/mock magic.
>
> I'm sorry, but I need today to focus in finishing patch for 3.6.7 bug, so
> I don't have time to investigate the issue, but IMO it needs to be resolved
> ASAP, otherwise
> more broken patches can be merged.
>
> @Eyal, is there somebody in CI team who can investigate this (assuming CI
> team members have much better knowledge of CI/Jenkins/mock magic
> and standard engine build command raised error correctly)?
>
> Thanks
>
> Martin
>
> [2]
https://gerrit.ovirt.org/58242
>
>
>> The loges here seem to do the same things:
>>
>>
http://jenkins.ovirt.org/job/ovirt-engine_master_check-merged-el7-x86_64/...
>> (See the check_merged.sh log)
>>
>> --
>> Barak Korren
>> bkorren(a)redhat.com
>> RHEV-CI Team
>>
>
>
--
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)