Open Stack

On Fri, 24 Jan 2014 11:58:42 -0800
Franck Yelles <franck110 at gmail.com> wrote:
> Hi everyone,
> 

Hi - note that the openstack-qa mailing list is no longer being used
(we should really remove it!). I've CC'd your email along to
openstack-dev with the [qa] tag in the subject line.

> I would need some clarification of the Tempest testcases.
> I am trying to run tempest on a vanilla devstack environment.
> 
> My localrc file has the API_RATE_LIMIT set to false.
> This is the only modification that I have.
> 
> I would run ./stack.sh and then run ./run_tempest.sh and would have X
> errors. (running the  failing testcases manually works)
> Then I would unstack and stack again and run again ./run_tempest.sh
> and would have Y errors.
> 
> My VM has 2 dual quad core and 8Go of RAM
> 
> Why do I have this inconsistency ? Or am I doing something wrong ?

So you'll probably need to analyse the failures themselves to see what
is going on. Perhaps post some of the output to a pastebin so we can
look at it and make the log files available. Asking on #openstack-qa 
might get you some help as well.

This sort of inconsistency in test results I've found is often due to
insufficient memory (although 8Gb should be enough unless you are
running other things at the same time) and/or running too many tests in
parallel and hitting some of the resource quota limits. But there should
be some clues to this in the log files.

Regards,

Chris