Instead of requiring manual understanding of how to generate
the states diagram, add an explicit virtualenv that can be
used to create and generate that docs diagram (and replace the
existing file with it).
Change-Id: I1738da9bf08bfda9b9be01489e8456569a31709e
Disabling erase_devices clean_step in agent_ilo driver doesn't work
even though the configuration option is updated to 0.
Change-Id: I72501793b9a0d384555bec2bbe2ddfdb2f1c2c2f
Closes-Bug:#1448103
This patch just add some DEBUG and INFO logs for the iSCSI deployment
methodology, so that the operator knows what's the steps the ramdisk is
taken by looking at the conductor logs.
The patch add logs for Ironic when working with both supported ramdisks
DIB and IPA.
Change-Id: Iaee09643b03877a1a735b6019def47445bf3e16f
Added autospec=True or spec_set= to mocks
Fixed one incorrect call to has_calls() and changed to assert_has_calls()
Change-Id: I01756a4db88e50a8a663a164644feebd9be90f2c
Create third_party_driver_mock_specs.py with an initial mock spec for
the pywsman 3rd party library.
Enable initial use of this mock spec.
Change-Id: I387d26ef490833f5e833ca214dbce009c5edff0d
In order to use iPXE with neutron we need to create a tag in Neutron,
this patch document how to do it.
Change-Id: I313eb3600e864ff7a868154fe2d0aa6fb8cc69a2
The API should not need to reboot for a neutron config option.
Follow up to 03d9dec4e4d94b598d54dd6a0d1b22d8be40b597.
Change-Id: Ic6462f31cef49fa943046b963989680240f292cf
In juno, ironic provided support for migrating from nova's
"baremetal" driver to its "ironic" driver. These scripts
were used for that.
The baremetal driver was removed from nova in kilo. Ironic
doesn't support migrating from pre-kilo nova's baremetal
driver to post-juno nova's ironic driver.
Hence, deleting these migration scripts.
Closes-Bug: #1446330
Change-Id: Ic4990caa83a2eca3ee9fbd105bfe0a12bdaa7efe
is_glance_image() returned True if image_href was a normal file like
"deploy.iso". It should return False for this case.
This patch fixed is_glance_image() predicate logic so that it returns
True only if image_href is glance URL or glance UUID.
Change-Id: I338964d05280175dd86bb2b6e1c7073871382277
Closes-Bug: #1446052
Ironic requires a cleaning network uuid in Kilo when using the default
config. This adds docs for how to configure the option and notes this
must be done when upgrading from Juno to Kilo.
Change-Id: Ia5cc118daefa7e3d966f1ff219ab024cb47d81f5
This removes the deprecated ironic compute driver and scheduler
manager. The compute driver and scheduler manager have been in
nova's tree since Juno and this was to have been deprecated in
the Kilo release. It is now Liberty, so it is about time :)
DocImpact
Change-Id: I8c3c35ca5a9a163e689a4c856b1efd557779b9fb
Add spec= and autospec=True to mocks
Fixed issues with misspelled mock asserts discovered because of this
work.
Example errors
assert_called_ince_with()
called_once_with()
Change-Id: Ie883399427e8e1066da1e166e3a8dbf546e57bee
So that the openstack manuals are consistent and use the
same (right) term, this replaces the term 'metrics' with
'meters' in the help for the config option
send_sensor_data_interval.
DocImpact
Closes-bug: #1436371
Change-Id: Id88058bf2fe070a8e0774857f2c5690a60c30de0
The agent returns command_result of None when a command is in progress.
The code assumed it would return an empty dict. Fix the access of
command_result to never return None, and fix the tests to reflect
reality.
Closes-Bug: #1444715
Change-Id: Iae6cd90517cb7b268ce2c089c92f64b09a606a42
Bump pre-version in setup.cfg to formally open Liberty development.
Kilo release branch will be cut from the previous commit.
Change-Id: I869afdbe3eb4f80277082329cb9a5794812f15f2
This patch is adding a note to the local boot documentation about it
being dependent upon a updated deploy ramdisk.
Change-Id: I16904cd1c9c942239c1416acbe3d8d0ce490d8f2
The agent driver is using RPCs to call back from the driver to the
conductor asynchronously. When using the RPC.call() method, some nodes
would end up with stuck locks when using the agent driver during cleaning.
The agent driver would issue a call() to continue_node_cleaning() after
either the first heartbeat (from prepare_cleaning) or a heartbeat after
a clean step had completed. The conductor would attempt to get a lock,
but would not be able to. The node would retain its locked state (so
far as I could tell), even after the error. Other nodes would continue
and complete cleaning just fine. The exception raised by
continue_node_cleaning() was likely not caught by the agent driver, but
caught by vendor_passthru() in the conductor as an expected exception.
Switching to cast() avoids the issue because the errors are not sent
back to the caller. I didn't experience any more stuck locks with
this change.
Change-Id: I4dbb04ccb93199bba4e1a1614bc19b70a068a9ea
Closes-Bug: 1442810
