Is this still an issue?

Interesting, in this particular case it doesn't look like a DNS issue. The connection was as follows:
wget -> proxy01.phx.ovirt.org -> jenkins.ovirt.org

Wget was able to access the proxy and the proxy tried to fetch the URL but got a 404 error from jenkins:

1503475892.751 4 66.187.230.21 TCP_MISS/404 878 GET http://jenkins.ovirt.org/job/ovirt-node-ng_master_build-artifacts-el7-x86_64//artifact/exported-artifacts/ - HIER_DIRECT/66.187.230.92 text/html

I get a 404 error when accessing this URL as well so there's some error in the fetching script:
http://jenkins.ovirt.org/job/ovirt-node-ng_master_build-artifacts-el7-x86_64//artifact/exported-artifacts/

why are we not appending the artifact name or build number to the URL above? In squid logs I see a successful fetch of the Appliance via the proxy with a correct URL:

1503475797.567 9880 66.187.230.21 TCP_REFRESH_UNMODIFIED/200 822060496 GET http://jenkins.ovirt.org/job/ovirt-appliance_master_build-artifacts-el7-x86_64/524/artifact/exported-artifacts/oVirt-Engine-Appliance-CentOS-x86_64-7-20170822.ova - HIER_DIRECT/66.187.230.92 application/octet-stream

This one has both a build number and file name and should be added here.

There are some occasional flaky DNS stuff going on. But this job has other problems since DNS is not that flaky.

is this still happening, do we know if there was a DNS outage?

It still uses lastSuccessful, but you are right , I added a call to buildNumber to see which images we're fetching. Since buildNumber is empty in this case, it produces a wrong download link. Question is, how can buildNumber be empty ?

We seems to had some DNS outage. Two change queue jobs failed at that time
with symptoms that look like DNS problem, e.g.:

2017-08-23 07:16:59,037::INFO::repoman.common.repo::Resolving artifact
source
http://jenkins.ovirt.org/job/vdsm_master_build-artifacts-el7-ppc64le/2361/
requests.exceptions.ConnectionError: ('Connection aborted.', gaierror(-2,
'Name or service not known'))

or

[ ERROR ] Yum Cannot queue package ovirt-engine-setup: Cannot find a valid
baseurl for repo: base/7/x86_64
[ INFO ] Yum Performing yum transaction rollback

I checked DNS resolution on jenkins master at least and it looks working
now.

Still might be specific to some particular environment/hosts, so will watch
it a bit more.

Anton.

–
Anton Marchukov
Team Lead - Release Management - RHV DevOps - Red Hat

I think the link generated by the suite is just incorrect:

should be:

Or whatever was agreed for the node suite.

Maybe its the same issue we see when failing on accessing gerrit.ovirt.org?
it might be global DNS issues.
Also, are we still using proxy.phx.ovirt.org?

2 Resolving proxy01.phx.ovirt.org (proxy01.phx.ovirt.org)... 66.187.230.40
08:11:32 Connecting to proxy01.phx.ovirt.org
(proxy01.phx.ovirt.org)|66.187.230.40|:3128...
connected.
08:11:32 Proxy request sent, awaiting response... 404 Not Found
08:11:32 2017-08-23 08:11:32 ERROR 404: Not Found.
08:11:32
08:11:32 + res=8

On Wed, Aug 23, 2017 at 11:43 AM, Yedidyah Bar David (oVirt JIRA) <

–

Eyal edri

ASSOCIATE MANAGER

RHV DevOps

EMEA VIRTUALIZATION R&D

Red Hat EMEA <https://www.redhat.com/>
<https://red.ht/sig> TRIED. TESTED. TRUSTED. <https://redhat.com/trusted>
phone: +972-9-7692018
irc: eedri (on #tlv #rhev-dev #rhev-integ)

The failure reason is:

Looks like it is missing the build number from the Jenkins link(iirc it was 'lastSuccessful' in the past), , I think you did some changes there a few weeks ago?