Ouptut for beakerlib tests

[Tim Flink]

There's a bug currently open for the standard test roles where there
are duplicate output lines in the logfile output when the beakerlib
role is complete.

https://pagure.io/standard-test-roles/issue/83

The summary of the issue is this: beakerlib outputs to stdout, stderr
and a logfile which do have some duplicate information between them but
none of them are identical. As a reference to what other beakerlib
environments do, Beaker seems to put all three into separate files.

What do we want to have as the primary output of the beakerlib role?

My instinct is to use the logfile as the primary output because it is
the easiest to scan, even if it is lacking details. STDOUT and STDERR
could then be combined into a second file or kept separate yet still
included in the artifacts generated.

Any thoughts on this plan or another method for getting duplicated
lines out of the log output?


Thanks,

Tim

[David Moreau Simard]

Hey,

We discussed on IRC this morning but writing here for posterity...

I saw that you were using Ansible -- I'm the author of ARA [1], an open source native
Ansible callback plugin that provides a CLI, web reporting interface and soon, a full REST
API [2].
I've been working on the Fedora packaging for ARA and it's available for rawhide
already [3].

I'm happy to help you get started if this sounds like it could be useful for your use
case, feel free to reach out to me on IRC (dmsimard) -- we also have an IRC channel on
freenode: #ara.

Let me know if you have any questions !

[1]: https://github.com/openstack/ara
[2]: https://dmsimard.com/2017/11/22/status-update-ara-1.0/
[3]: https://koji.fedoraproject.org/koji/packageinfo?packageID=24394