Add 'dry-run' functionality to the test runner
prashantmital opened this issue · 1 comments
Users should be able to set a flag to use astrolabe in dry-run mode to make it faster to debug issues when integrating new drivers. Integrating a new driver usually requires running the test suite multiple times in quick succession with changes being made to the evergreen configuration and/or the driver-provided scripts. Currently, this process is extremely cumbersome as some maintenance tasks can take a very long time so users must wait for a long time before being able to see the outcome of a given run.
In dry-run mode, astrolabe would simply run the driver workload for a pre-determined duration of time on an existing cluster. No maintenance would be performed. The cluster would not be torn down at the end of the run. An additional flag could be used to specify which cluster to use or one could be selected at random from what is available.
#35 adds a mechanism for validating the workload executor which makes it unnecessary to do this work.