Note that I am not questioning the
intent, but am worried about the results as-currently-implemented.
It is very important to understand what artifacts will get into
that distribution, since if that image will undergo CSIT to see
whether a controller change did not break OF, that OF piece needs
to be self-consistent.
Jenkins are just like threads, with Nexus being memory. What I am
asking is: what are the locks the reader (controller-distribution
job) holds to prevent writers (all the merge and integration jobs)
from modifying state (artifacts in Nexus) as it is being read to
assemble the distribution?
If we have no such assurances, what assurance do we have in the
reliability and repeatability departments? When things break, how
do we track which source code corresponds to that particular
controller-build?
Thanks,
Robert
On 02/26/2015 08:09 PM, Luis Gomez wrote:
toggle quoted message
Show quoted text
I am not sure I follow here, can you explain more:
1) false positives, from where?
2) high-impossible to track down, the only artifact
that really changes is the controller artifact, is that so
difficult to figure out?
BR/Luis
On Feb 26, 2015, at 11:03 AM, Robert Varga < nite@...>
wrote:
On 02/26/2015 07:53 PM, Luis
Gomez wrote:
Hi Robert, see my comments in-line:
On Feb 26, 2015, at 10:33 AM, Robert
Varga < nite@...>
wrote:
So the problem I
see is the following:
1) a controller committer merges a patch
2) controller-merge-master publishes artifacts
3a) controller-distribution triggers
3b) openflowjava-integration-master triggers
Now we have two jobs running (3a, 3b) and
based on the layout of job queue, slave
spinup, etc. etc., three things can happen:
- 3a may pick up artifacts from 3b, or
- 3a picks up
artifacts before 3b, or
- 3a picks up some artifacts before 3b and
some after that
The only requirement for 3a) is to pick
new controller artifacts, whether or not we use old
or new artifacts from openflowjava is not that
important IMO if we consider the integration job
will not modify much the existing project artifacts
(i.e. no project code change) unless there is a
change in a 3rd party version in odlparent, I think
this last was the reason we added the publish step
in the integration job.
Yeah, but the point is that anything that uses that
distribution may end up with an inconsistent image, so you
may end up with various false positives. And those will be
transient and nigh-impossible to track down.
Bye,
Robert
The same thing is going to happen if I quickly
merge a bunch of patches...
Bye,
Robert
On 02/26/2015 06:56 PM, Luis Gomez wrote:
For some reason I neglected this
mail, see my answers in-line:
On Feb 25, 2015, at 1:00 AM,
Robert Varga < nite@...>
wrote:
Hello,
I am not sure I follow what exactly
is the plan here. Is there going to
be a centralized 'this is what
integration looks like' job?
Furthermore how does this interact
with the ${project}-integration
jobs, which also publish artifacts,
but are not configured to trigger
according to the dependency chain?
${project}-integration
builds the project code on upstream
merges, ${project}-distribution builds
the integration project on own project
merges so they are very different
So I
merge stuff to yangtools, which
triggers a merge job, which will
complete. Now two jobs will trigger:
the controller integration job and
this distribution job ... controller
integration job publishes artifacts,
so ... which artifacts is the
distribution job going to pick up?
integration artifacts = karaf
distribution
Bye,
Robert
On 02/25/2015 03:20 AM, Luis Gomez
wrote:
Hi all,
I am pushing following
patches to releng repo:
This patch will create
a job that downloads and builds
the integration repository
whenever there is a merge in the
above projects. You can check the
job template here:
You will need this job
for 3 reasons:
1) To verify your
changes are compatible with the
integration code
2) To make sure your
changes are immediately included
in the integration distribution
3) To run some system
test after the distribution is
built
Please let me know if
you have questions or concerns
about merging this.
PS - these jobs will
replace the old
integration-polling jobs +
integration-centralized-integration
job
BR/Luis
_______________________________________________
integration-dev mailing list
integration-dev@...
https://lists.opendaylight.org/mailman/listinfo/integration-dev
|
