nrsysmond suite fails due to invalid license key, but only on some platforms

[cwjohnston]

I am seeing failures from nrsysmond-default test suite which appear to be related to the TESTKEY_NRSYSMOND key being invalid. For some reason I can consistently reproduce this on certain platforms (e.g. centos-59, ubuntu-1204), but I believe it is actually a timing issue and not particular to the platforms in question.

[chr4]

I encountered the same problem. as TESTKEY_NRSYSMOND is obviously not a valid newrelic key, this is expected.
On some platforms, nrsysmond apparently doesn't start up when the key is not valid, while on others it comes up.

I tried the test with our newrelic key, then it passes.
Unfortunately, I couldn't find a valid testkey for newrelic, and a random shasum of the same length doesn't work either. (I tried 4e1243bd22c66e76c2ba9eddc1f91394e57f9f83, the shasum of "test\n")

A workaround would be to not start the nrsysmond service when testing, but I have no idea howto implement that cleanly.

[cwjohnston]

I will work with my contact at New Relic and see if there is a valid test key we can use. Otherwise, I just wanted to keep track of this issue even if there's not much to be done about it at this point.

[chr4]

A testkey would be great!
I agree, tracking issues is a good thing 👍

[jeffbyrnes]

👍 to having a test key.

[maxim]

I once wrote to their customer support asking for something like fake key for testing, they said that they don't have anything like that. @cwjohnston, was hoping you may've had better luck. Any news on this?

[cwjohnston]

@maxim I didn't hear any thing back on the request I sent to NewRelic, and, as I am no longer working for the company where I was using their service, I don't expect to receive a response from them.

[ashew]

A bit of a necro 2 years later, but I've had success with '1234567890123456789012345678901234567890' - if it's someone legit API key, then I'm truly sorry (and jealous - gotta love sequential license keys!).

[chr4]

Can someone verify that this solves the issue with rhel/ ancient ubuntu?

[chr4]

To be honest, I'm not quite sure whether settings this is a good idea. If you just install the cookbook without changing the attribute, this would lead to potentially confidential data being send to a potentially random newrelic user.
Can someone confirm that this key can be used for testing? Alternatively, we could raise if the key is not changed.

[jeffbyrnes]

@chr4 rather than have it be the default, why not just use it as a Test Kitchen attribute, and leave the actual attribute blank, and error out if it’s not set?

[chr4]

This sounds like a good idea.

[chr4]

The latest commit on master uses 1234567890123456789012345678901234567890 as a key in th tests.