Description of problem: When trying to delete network agent openstack command fail with exception. Using neutron command for the same agent-delete pass with no errors. "[stack@undercloud-0 ~]$ openstack network agent delete b0fdb309-15da-4f77-824a-ed91db451907 Failed to delete network agent with ID 'b0fdb309-15da-4f77-824a-ed91db451907': get_agent() got an unexpected keyword argument 'ignore_missing' 1 of 1 network agents failed to delete. [stack@undercloud-0 ~]$ " Version-Release number of selected component (if applicable): How reproducible: try to delete network agent. Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
The package version is: python-openstackclient.noarch 3.8.1-0.20170208195351.1e3dc48.el7ost
Can you say why do you consider this a release blocker? Is agent-delete used by Director somehow? This doesn't fit blocker criteria as a workaround exists by definition...
Hi, I have consider this as release blocker as agent-delete is used in the scale down and compute node replace procedure.
So, as of build 2017-04-24.2 this issue still exists and it blocks scaling down for split-stack RFE. Are we planning to do anything with it?
Hi Jakub Are we going to get a fix for this in osp 11? If not this this needs to be documented. Thanks
It seems like a problem with unified CLI command and not with a component, "neutron agent-delete" should work
Triaged for current release.
*** Bug 1482506 has been marked as a duplicate of this bug. ***
Hi. Could you clarify what documentation needs to be updated? Do I understand correctly that this only affects split-stack use cases, or is it a problem we need to document the workaround for elsewhere as well?
(In reply to Julie Pichon from comment #14) > Hi. Could you clarify what documentation needs to be updated? Do I > understand correctly that this only affects split-stack use cases, or is it > a problem we need to document the workaround for elsewhere as well? No, it's not specific to split stack, starting RHOS 11 "openstack network agent delete <uuid>" does not work, but required for Node removal/replacement procedures as per official documentation. There is old command, that is mentioned as deprecated, but still works: "neutron agent-delete <uuid>", however there is no mentioning of that anywhere in a documentation. I'm afraid that people who are unfamiliar with old-style commands like neutron does not know how to work-around this issue, e.g. it needs to be documented until issue is resolved. Since it was not fixed neither in RHOS 11 nor in RHOS 12 we have 2 affected versions right now.
Do we want to backport it back to RHOS 11?
Yes, I'm going to propose a backport to Ocata/OSP11 as well. Just want to run a couple of manual tests first.
(In reply to Julie Pichon from comment #17) > Yes, I'm going to propose a backport to Ocata/OSP11 as well. Just want to > run a couple of manual tests first. Perfect, thank you
*** Bug 1533268 has been marked as a duplicate of this bug. ***
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2018:2521