When parallels creates linked clones, it generates a new snapshot on the
cloned VM with the name 'Snapshot for linked clone', which when used
with jenkins results in hundreds of unneeded snapshots per week.
Trying to assert that an exception was raised using
helper_open.write.assertRaises() is bogus--there is no such method. Use
standard unittest.assertRaises() instead.
Otherwise we are not properly mocking this function call and the
tests fail when the `_conn3` function is called. (Any distro using
salt.utils.boto3 will fail here.)
We also need to protect the test class if the correct versions of
boto and moto are not installed on the test system.
* Change osmajorrelease back to str in test_useradd
* Change osmajorrelease back to str in test_userdel
* Change osmajorrelease back to str in test_members
* Openstack keystone V3 API support in keystone modules.
* Python3 compatibility unicode to six.text_type
* Return values of tests: test_user_absent, test_endpoint_absent,
test_role_absent, test_service_absent, test_tenant_absent
* Fixed tests states.test_role_present and states.service_present
* Fixed module tests (not all)
* These are the old tests and need to be rewritten anyway
* fix pylint error
* fix keystone state tests
* Make python3 and python26 happy
* Don't need both ways of nesting with statements
* No longer need pylint disable
* Lint fix
* take out the correct import
- The previous implementation was limited to the 15 char limit for "Computernames" which would cause the state to always fail if the hostname was longer
* Don't test for linux shells on Windows
* Don't force windows to be false
Salt didn't used to support runas on Windows.
Salt now supports runas on Windows. This line was causing the wrong code
paths to be followed.
* Skip test on windows because no VT on windows
* add missing closing paren
Also load the __utils__ into pillar and fileserver and fileclient so
that this function can be used there with __opts__.
Also sets the default for s3 to 'us-east-1' like it is documented.
Closes#30976Closes#28913
* This test was doing all sorts of things, raging from simply failing
to driving CPUs to 100%, to leaking memory until the entire test suite crashes.
This seems to happen only in conjunction with running the entire unit test suite
and does not occur on its own
cc: @kraney
* Better method to skip test. This way we can track it.
* Remove the deprecated version of the status.uptime function.
The deprecated version of the function was set to be removed in
Carbon. The Carbon release notes were also updated with this change.
* Remove test_deprecated_uptime unit test
The deprecated version of uptime no longer exists.
* Remove unused import and deprecated uptime reference
to driving CPUs to 100%, to leaking memory until the entire test suite crashes.
This seems to happen only in conjunction with running the entire unit test suite
and does not occur on its own
cc: @kraney
