Debugging across openstack services, specially network or storage issues, is becoming increasingly difficult due to the distributed nature of its services and agents.
In this talk we plan to present a methodology for debugging live deployments and development systems as well.
We will present the oslogmerger tool, which is capable of retrieving and aggregating logs across several hosts and several services and coalescing the logs into a single and meaningful linear one, making cross project/cross agents calls much more easy to follow.
The attendees should expect to learn a simple methodology based in ansible and oslogmerger to reproduce production issues, gather specific logs, augment them with system probes that will be mixed into the logs themselves.
Such reproducer ansible scripts help in many ways:
* developers can leverage them to reproduce them and generate bugfixes
* the operator can verify if a bugfix correctly fixed the issue.