Add integration test for proxy #1029
Conversation
danielmitterdorfer
added
enhancement
|
@elasticmachine test this please |
|
@elasticmachine test this please |
| function kill_related_es_processes { | ||
| # kill all lingering Rally instances that might still be hanging | ||
| set +e | ||
| # kill all lingering Elasticsearch Docker containers launched by Rally |
There was a problem hiding this comment.
I see that you removed the portion of this method that kills any rally pids found using jps.. Do we not need this functionality anymore w/ the release tests?
There was a problem hiding this comment.
Looking at this more, I assume this is because we are only concerned w/ cleanup from the script itself, and since the test is only running docker-compose test commands, we should have no need to kill rally outside of docker.
There was a problem hiding this comment.
Correct. There are only Docker processes so no need to kill anything else.
| cfg = it.ConfigFile(config_name=None) | ||
| log_file = os.path.join(cfg.rally_home, "logs", "rally.log") | ||
|
|
||
| if os.path.exists(log_file): |
There was a problem hiding this comment.
just curious, why would we not just zero out the log contents if we want a new logfile every time, instead of moving its contents. I assume the contents of the "last run" are useless if its a passing test by the time you get to the next test.
There was a problem hiding this comment.
The idea behind this is that we should keep the full history of the log file. I am not so concerned with CI but rather when you run the tests locally. It might be puzzling (and maybe even a problem) if we start truncating the log file.
|
@hub-cap can you please have another look at this? |
|
Thanks for the review! |
With this commit we migrate the integration test for HTTP proxy connections to
the new integration test infrastructure and cleanup any remaining leftovers in
the old script (which is still used for Docker release image tests).
Relates #975