Have remote tests scan all facts. #41
Conversation
| known_facts = host['facts'] | ||
| break | ||
| for fact in known_facts.keys(): | ||
| try: | ||
| assert (str(known_facts[fact]) in str(row[fact])) | ||
| assert (str(known_facts[fact]) in str(row.get(fact))) |
There was a problem hiding this comment.
One observation is that if the row does not have the fact then it will convert to the string 'None' which won't match. But if the code get updated it may break
>>> str(None)
'None'There was a problem hiding this comment.
Also why not use == here? It could have the false positive of 'abc' being in 'abcd'
There was a problem hiding this comment.
I was thinking that this would be more resilient to white space changes, when I was originally writing this and hand encoding the config file I think I had some issues with that.
There was a problem hiding this comment.
As per the 'None' issue, am going to make a change and push it and we can see what you think.
|
Now I check for facts not present in the scan and register errors if there are expected facts missing from actual scan. This should circumvent the problem around converting possible |
| known_facts = host['facts'] | ||
| break | ||
| for fact in known_facts.keys(): | ||
| missed_facts = known_facts.keys() - row.keys() | ||
| if (missed_facts != set()): |
There was a problem hiding this comment.
Empty lists means False for python:
>>> bool(set())
False
>>> bool(set([1,2,3]))
True
With that we can write cleaner code like:
if missed_facts:
...
| if (missed_facts != set()): | ||
| # This means there are facts that we expected, | ||
| # but were not collected | ||
| for fact in missed_facts: |
There was a problem hiding this comment.
The above if block is not needed the for can be used directly. If the missed_facts is empty no iteration will run.
| try: | ||
| assert (str(known_facts[fact]) in str(row[fact])) | ||
| if fact == 'cpu.bogomips': |
There was a problem hiding this comment.
Better to move the if outside the try/catch that is something that won't raise the expected issue and will help others understand what you are really trying to catch here.
| fact, | ||
| row.get(fact), | ||
| known_facts[fact], | ||
| )) | ||
| scan_errors.append(msg) | ||
| if len(scan_errors) != 0: |
There was a problem hiding this comment.
Not on this PR but this line is the same as if scan_errors:
Also, fixes problem of key errors if expected fact not present in scan, will fail more gracefully with informative error message now. Closes #40
1 similar comment
Also, fixes problem of key errors if expected fact not present in scan,
will fail more gracefully with informative error message now.
Closes #40
I ran the remote tests against the vcenter machines as well as against my local machine, and this fixed the key jboss fact key errors I was getting!
Once this is merged I will update jenkins config files so it will run tests against all facts.