Closed Bug 1019929 Opened 10 years ago Closed 10 years ago

Running 'make test' results in KeyError on Loop-Server Stage and Prod

Categories

(Hello (Loop) :: Server, defect)

defect
Not set
normal

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: jbonacci, Unassigned)

References

Details

(Whiteboard: [qa+])

Installed the very latest to get support for 'make test', 'make bench', and 'make megabench'.

Running the following command:
make test SERVER_URL=https://loop.stage.mozaws.net

Results in the following error:
1 occurrences of: 
    KeyError: 'hawk-session-token'    Traceback: 
  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/unittest/case.py", line 331, in run
    testMethod()
  File "loadtest.py", line 17, in test_all
    self.register()
  File "loadtest.py", line 36, in register
    resp.headers['hawk-session-token'])
  File "/Users/James/loop-server/loadtests/venv/lib/python2.7/site-packages/requests/structures.py", line 77, in __getitem__
    return self._store[key.lower()][1]


For comparison, running the most simple 'make test' (which defaults to Stage with port :443) gets the same error.
Whiteboard: [qa+]
Just verified on Prod after a much longer delay hitting Prod vs. Stage.
Summary: Running 'make test' results in KeyError on Loop-Server Stage → Running 'make test' results in KeyError on Loop-Server Stage and Prod
Blocks: 994961
This seems to be because the server wasn't up to date with the script (e.g. it wasn't returning hawk sessions).

I've tested it manually and just had a local run. Closing this.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Working great on Stage.
Still a bug on Production. 

Are there differences between the two environments right now that prevents a successful test?
FYI: QA will not often be hitting Production with this test unless we want to do that going forward....
Yes, that's because production needs to be updated to latest master (or the 0.4 tag) and that's not the case yet.
Which actually remembers me we need to define a process to deploy our different environments. James, can you provide some feedback about how you do that for the other services we have?
We can open a new ticket for process if you want, but in short, OPs keeps a number of Mana pages going plus issues in svcops and puppet-config repos to define the Stage and Production deployments for the various services...
Status: RESOLVED → VERIFIED
You need to log in before you can comment on or make changes to this bug.