WIP ContainerInfra: extract ID from cluster resize result

[tghartland]

The cluster resize result should contain just the cluster's ID, so it needs its own Extract method. Currently magnum is sending an empty response which is causing a different error in gophercloud, this will be fixed in magnum with [1].

Will keep as WIP until the fix is merged in magnum and I have been able to test with it.

For #1631

Links to the line numbers/files in the OpenStack source code that support the
code in this PR:

[1] https://review.opendev.org/#/c/672600/

[coveralls]

Coverage increased (+0.007%) to 76.997% when pulling 5e40600 on tghartland:magnum-resize-fix-1631 into c99da27 on gophercloud:master.

[theopenlab-ci]

Build succeeded.

• gophercloud-unittest : SUCCESS in 3m 09s
• gophercloud-acceptance-test : SUCCESS in 2h 07m 35s
• gophercloud-acceptance-test-ironic : SUCCESS in 26m 28s

[jtopjian]

recheck stable/newton

[jtopjian]

oops - wrong PR. Ignore those test results :)

[theopenlab-ci]

Build failed.

• gophercloud-acceptance-test-newton : FAILURE in 53m 12s

[jtopjian]

@tghartland It looks like the upstream patch has been merged - is it ready for review? If so, can you confirm that the UUID is the only field being returned now (you're removing the node_count field in the test fixtures).

[tghartland]

@jtopjian Yes the fix is merged upstream, I wanted to wait until we had it deployed here so I could test this change and be 100% certain it works (but it's taking longer than expected to get the magnum fix deployed).

This can be reviewed now and I can test later.

Since this changes the results API for extracting the ResizeResult, is there anything special that has to be done or should it just be documented whenever the next release is made?

[jtopjian]

@tghartland I'm totally fine waiting until you're able to confirm functionality. We can then also verify if the response body has changed.

If the response has changed, we might want to change how the new extraction is done, similar to what's being done here.

[tghartland]

@jtopjian the response should just be the UUID of the cluster

{
   "uuid":"746e779a-751a-456b-a3e9-c883d734946f"
}

The code in this PR should handle that already, is the code you linked doing something else?

[jtopjian]

Ah, yeah, I see what's going on.

Prior to this PR, the result was being extracted into a generic Cluster struct. This PR implements a more fine-tuned extraction of just the UUID. So it's quite possible that node_count was never returned in the first place.

Since the results of the extraction method are being changed, we'll want to document it - I'll do that in the changelog.

But again, this can all wait until you've confirmed functionality.

[theopenlab-ci]

Build failed.

• gophercloud-unittest : SUCCESS in 3m 14s
• gophercloud-acceptance-test : FAILURE in 12m 38s
• gophercloud-acceptance-test-ironic : FAILURE in 12m 51s

[tghartland]

@jtopjian sorry it's been so long with no activity here. I've now tested this on the magnum master branch and it does work. (The fix has been released in magnum Train now as well).

[jtopjian]

@tghartland No problem at all. I'm going to go ahead and merge this. Since Magnum isn't being actively tested in OpenLab, the test results won't really matter.

Thank you for your work on this.