automatically use novaclient plugins

[gtmanfred]

no more having to specify what module the extension should overwrite

https://pypi.python.org/pypi?:action=search&term=_python_novaclient_ext&submit=search

[ghost]

Test Failed.

If the failures are unrelated to your code, don't stress, a core developer will know these apart.
In the future, if possible, please branch off a passing commit to avoid false positives.
Refer to this link for build results: http://jenkins.saltstack.com/job/salt-pr-build/7128/

[ghost]

Test Failed.

If the failures are unrelated to your code, don't stress, a core developer will know these apart.
In the future, if possible, please branch off a passing commit to avoid false positives.
Refer to this link for build results: http://jenkins.saltstack.com/job/salt-pr-build/7129/

[ghost]

Test Failed.

If the failures are unrelated to your code, don't stress, a core developer will know these apart.
In the future, if possible, please branch off a passing commit to avoid false positives.
Refer to this link for build results: http://jenkins.saltstack.com/job/salt-pr-build/7130/

[ghost]

Test Failed.

If the failures are unrelated to your code, don't stress, a core developer will know these apart.
In the future, if possible, please branch off a passing commit to avoid false positives.
Refer to this link for build results: http://jenkins.saltstack.com/job/salt-pr-build/7131/

[gtmanfred]

@techhat @thatch45 i would really like to see this get merged back into 2014.7 if at all possible, since this is a new driver for 2014.7 anyway.

[rallytime]

Can we hold of on backporting this, please? There have been some very fishy test-failures on the develop branch and they're pointing to this PR merge. Haven't figured out why yet, but @cachedout and I are looking into it.

[gtmanfred]

https://github.com/saltstack/salt/pull/15120/files#diff-87e66d53ad1a09e0e96e344a932839f3L421

Check this change, this would be my guess as to why it might be failling.

[rallytime]

@gtmanfred - Yeah, you're right. This is breaking our tests with OSError: [Errno 12] Cannot allocate memory errors all over the place. Is there a more efficient way to implement this?

[gtmanfred]

I know what the problem is, but I don't know a good solution... give me a bit.

It is adding the modify_body_for_create hook a ton of times and probably causing a memory leak cause it never checks if that function was added or not, this really should be fixed by upstream, but I am going to see if I can fix it.

[gtmanfred]

http://ix.io/dZ8

I think this will fix it, it will make sure that the extensions don't continue getting loaded and cause it to cascade out of control.

[thatch45]

I am going to revert these because it is not fixing it. We will need to just take this one a little more slowly

[gtmanfred]

I blame novaclient for this... they did extensions the worst possible way.

Did yall get any idea what was actually taking up all the memory?