[automated-testing] how does glusto based testing handle logging?

[Sankarshan Mukhopadhyay]

On Mon, Apr 29, 2019 at 9:05 PM Jonathan Holloway <jholloway at redhat.com> wrote:

> In a nutshell, no. Glusto itself only sets up a main Python logger for the Glusto library to write to, typically when a test is run via /usr/bin/glusto.
> Above that, the gluster libraries either write to the main Glusto log or create their own logger for the tests themselves.
> For collection of those and logs at the system layer (syslog, Gluster, etc), the automation is handled at the Jenkins level.
>

I think this leads to a design decision around how a Glusto instance
for Gluster would need to have a centralized logging or, at the very
least, rsyslog configured. Logs resulting from failed tests are a good
place to start for patterns and identify remedial approaches.

> A centralized solution for test and log analysis is on the CCIT roadmap in QE, and I'm working with Data Hub for the Log File Analysis aspect using their Elastic.
> Framework (PyTest, Glusto, etc.) and test logs generated by the gluster-specific libraries would be included in logs collected by the system, but setup for collation of that log data would be a level higher.
>
> Someone can correct me, but I think the closest we have right now is the sosreport generated and stored after the test downstream.

I'm not sure who can respond to this.