Luis Gomez <luis.gomez@...>
toggle quoted messageShow quoted text
OK, I changed the test case order so that we start with the most basic services and end with most abstracted services :
Host Tracker & Simple Forwarding
Note that basic service does not necessarily means simple to test, so maybe I was not very precise when I said simple-to-complex. What I really meant was something
like basic-to-extra functions.
I also added the required steps on every area so every test case above is self-complete. Please review the test plan and let me know if you agree with it.
As for your question, yes, we can categorize the services in different ways like for example: basic network functions (Switch Mgr, Topology Mgr, FRM, Stats
Mgr), extra network functions (Host Tracker, Simple Forwarding, ARP Handler, Forward Mgr), basic node functions (Configuration, User Mgr, Connection Mgr) and extra node functions (Container Mgr, Cluster Mgr). This is just an idea but it could be more, anyway
besides the classification the important is that we do not leave features/modules without test.
From: Gmail [mailto:yangbaohua@...]
Sent: Friday, November 01, 2013 10:22 PM
To: Luis Gomez
Subject: Re: System Test Plan discussion
Sure, Luis. This is a valuable question!
IMHO, the simple-to-complex order is good.
However, we might also keep test cases independent from each other, i.e., each test case should be self-complete. Because, we sometime may want to test the function of individual module, instead of the entire platform.
Besides, we may even categorize the tested modules based on their functions, test complexity, etc. For example, state collection modules, basic forwarding modules, QoS modules.... And each category can be tested separately.
BTW, I think I would prefer some logic in the order like (simple to complex) and independent modules even it this means more TCs.
Hi Baohua and all,
I have also observed the list of TCs in the Base Test Plan has been reordered following alphabetical order. This is fine but I just want to explain the reason
for the order before:
1) I started from more simple (less abstraction) test cases modules like Switch Mgr or Toplogy Mgr, FRM and finished with more complex (more abstraction) ones
like: host tracker, ARP Handler, Container Mgr, etc.. The reason for this is if we start testing simple things and these fails, there is no point to continue with complex ones. Also if the test starts with a complex case and fails, more difficult to debug.
2) I also combined some modules like for example if I need to have some flows (FRM) in order to check statistics (stats Mgr), I create the flows (FRM1), then
I check the statistics (Stats Mgr) and then I clear the flows (FRM2). This way I do not need to create flows 2 times, one for FRM and another for Stats Mgr.
The thing is the existing TC flow does not work as the way I wrote it there are some modules (i.e. Stats Mgr) that needs to have certain conditions created
in previous modules (FRM). Anyway, instead of fixing this I would like to open a discussion (this is the right time) on how we want to write and later execute test plan:
1) Module order: do we want to follow any logical order (simple to complex or something else) or we just do alphabetical?
2) Module dependencies: do we want to have modules depending on the result of previous modules (more test efficient, less TCs) or have totally independent modules
(less efficient but very flexible if for example I want to run just one module test)
What do you think?
From: Baohua Yang [mailto:yangbaohua@...]
Sent: Friday, November 01, 2013 1:42 AM
To: Luis Gomez
Cc: integration-dev@...; huang denghui
Subject: Re: [integration-dev] Next steps for system test
Please help verify and input more details, or to see if we have better place to move to.
Denghui and I are working hard on the code writing these days, and we plan to release a workable one for the base edition in this week.
Currently, the simple list_nodes tests on arp handler, host tracker and switch manager have been provided.
Any one is welcome to make contribution to the code, documents or bug fixing.
On Fri, Nov 1, 2013 at 12:47 PM, Luis Gomez <luis.gomez@...> wrote:
It is been a busy day but finally I got some time to summarize our afternoon discussion around the system test:
- Phython scripts: Baohua will write some guide in the wiki on how to use the scripts. These can be very useful to debug test cases.
- Robot framework: It is already installed in Ericsson Lab, Carol will write some quick guide on how to use it. Denghui, you have also Robot experience so you can help Carol with this.
- TestON framework: We have meeting next week to know more about it (especially APIs and driver support), I will share with you some documentation I have already got from Swaraj. From the TSC call today, it looks
like we are interested in collaborating with ON.Lab people so this could be a good opportunity.
- System Test Plan: We need to continue working on this, everybody is invited to contribute in the wiki. I am going to meet Madhu next Monday to talk about OVSDB inclusion in base release, I will also work on
some test cases around OVSDB plugin. Punal is going to take a look on VTN project (Virtualization release) as they have all very well documented in the wiki. The rest of the projects we will need to ask information as we are writing the test plan.
So, the plan for next week is to get familiar with Robot framework and the python scripts created by China team, evaluate TestON framework, and continue filling the test plan.
integration-dev mailing list