On Wed, Mar 21, 2018 at 12:15 PM, Petr Horacek <phoracek@redhat.com> wrote:
I tried to retrigger the build several times, it was always executed on el7 machine, maybe it picks fc26 only when all other machines are taken?

Shouldn't be "Permission denied" problem detected in link_bond_test.py:setup_module()? It runs "check_sysfs_bond_permission".

I think that the error is from attempting to set a values that is not acceptable by the attribute, not because there is no permission to access sysfs.


2018-03-20 18:12 GMT+01:00 Edward Haas <ehaas@redhat.com>:
The tests ran on a fc26 slave and our bond option default map is in sync with the el7 kernel.
It looks like we MUST generate a bond default map on every run.

I'm a bit surprised it never happened until now, perhaps I'm not interpreting correctly the tests helperĀ  code? Petr?
Assuming I'm correct here, I'll try to post a fix.

Thanks,
Edy.

On Tue, Mar 20, 2018 at 12:14 PM, Dan Kenigsberg <danken@redhat.com> wrote:
+Petr

On Tue, Mar 20, 2018 at 11:07 AM, Francesco Romani <fromani@redhat.com> wrote:
> Hi all,
>
>
> we had a bogus failure on CI again, some network test failed and it
> seems totally unrelated to the patch being tested:
>
>
> http://jenkins.ovirt.org/job/vdsm_master_check-patch-el7-x86_64/22410/consoleFull
>
>
> could someone please have a look?
>
>
> Bests,
>
> --
> Francesco Romani
> Senior SW Eng., Virtualization R&D
> Red Hat
> IRC: fromani github: @fromanirh
>