When looking at sources in Integration/Test,

only suites in csit/suites/controller/dom_data_broker/

are sensitive to theprotocol used.


Suites with "pref" in their name

should be run with tell-based protocol.

Other suites in that directory could run

with both protocols, but frequently they would fail

on ask-based protocol as it does not handle those scenarios.


Other suites in csit/suites/controller/

should work with either protocol,

perhaps only cluster_singleton/ ones could

behave differently due to bugs.


See the table in docs/cluster/carbon/release.rst

for which combinations were important for us to track

(-a is ask-based, -t is tell-based;

better decoding at docs/cluster/carbon/tests.rst).


Vratko.


________________________________
From: Tom Pantelis <tompante...@gmail.com>
Sent: Friday, August 31, 2018 2:05:05 AM
To: Jamo Luhrsen
Cc: controller-dev; integration-...@lists.opendaylight.org; Vishal Thapar; 
Vratko Polák
Subject: Re: controller cluster csit suites for ask or tell or both



On Thu, Aug 30, 2018 at 5:44 PM, Jamo Luhrsen 
<jluhr...@gmail.com<mailto:jluhr...@gmail.com>> wrote:
Tom, Vratko, or any other expert,

Can you help us figure out this patch:

https://git.opendaylight.org/gerrit/#/c/74692/

The goal is to split the suites in to two jobs so we can only have
an ask based job and a tell based job. It seems that some of the
suites will work in both, but others are only intended for one or
the other.

This will give us better clarity as to the stability of each
protocol over time. I'm guessing this is going to require a little
digging in to each suite's robot .html files to try and read
what's going on.

ask based:
https://logs.opendaylight.org/releng/vex-yul-odl-jenkins-1/builder-copy-sandbox-logs/374/thapar-controller-csit-3node-clustering-ask-all-oxygen/1/robot-plugin/log.html.gz

tell based:
https://logs.opendaylight.org/releng/vex-yul-odl-jenkins-1/builder-copy-sandbox-logs/375/thapar-controller-csit-3node-clustering-tell-all-oxygen/1/robot-plugin/log.html.gz


>From 
>https://logs.opendaylight.org/releng/vex-yul-odl-jenkins-1/controller-csit-3node-clustering-all-fluorine/185/robot-plugin/log.html.gz,
> each test is preceded with "Restart Odl With Tell Based True/False", so I 
>think we can go by that pattern to start.



Thanks,
JamO

_______________________________________________
controller-dev mailing list
controller-dev@lists.opendaylight.org
https://lists.opendaylight.org/mailman/listinfo/controller-dev

Reply via email to