Created attachment 1202522 [details] entire evm.log in bz format Description of problem: Instance provisioning fails on RHOS9 The requested availability zone is not available Version-Release number of selected component (if applicable): 5.6.2.0 RHOS9 GA How reproducible: 100% Steps to Reproduce: 1.login to cfme ui 2. go to compute -> cloud -> instances 3. fill the network, instance name, tenant... Actual results: Request failure: Request ID 153,000,000,000,002 Status Error Request State Finished Requester Administrator Request Type VM Provision Description Provision from [cirros] to [ddd] Last Message [Excon::Error::BadRequest]: Expected([200, 202]) <=> Actual(400 Bad Request) excon.error.response :body => "{\"badRequest\": {\"message\": \"The requested availability zone is not available\", \"code\": 400}}" :cookies => [ ] :heade Expected results: provision success, the API call will arrive the openstack cloud provider Additional info: [----] I, [2016-09-19T08:41:53.308252 #15731:609988] INFO -- : Q-task_id([log_s tatus]) MIQ(MiqServer.log_system_status) [EVM Server (15538)] MiqQueue count for state=["dequeue"] by zone and role: {"default"=>{nil=>1, "smartstate"=>1}} [----] I, [2016-09-19T08:41:53.308325 #15731:609988] INFO -- : Q-task_id([log_s tatus]) MIQ(MiqServer.log_system_status) [EVM Server (15538)] MiqQueue count for state=["error"] by zone and role: {"default"=>{nil=>1}} [----] I, [2016-09-19T08:41:53.308382 #15731:609988] INFO -- : Q-task_id([log_s tatus]) MIQ(MiqServer.log_system_status) [EVM Server (15538)] MiqQueue count for
The error message is one returned from OpenStack, meaning the request is making it to the OpenStack API. I think we need to know more about the QE setup and environment, as this could be an OpenStack issue. In fact, I wonder if it has anything to do with: https://bugs.launchpad.net/fuel/+bug/1610309 ?
Mainn, the availability zones indeed missing on my RHOS setup (Mitaka) nova aggregate-list +----+------+-------------------+ | Id | Name | Availability Zone | +----+------+-------------------+ +----+------+-------------------+
Adding more interesting part of evm.log logfile [1]. It says, that instance should be provisioned with AZ named "volume-nova" which does not exist. I believe it is a UI on controller caused bug. "nova" availability zone should be present in OSP deployments by default IMO. [1] http://paste.openstack.org/show/581760/
It looks to me as a duplicate of https://bugzilla.redhat.com/show_bug.cgi?id=1373559 (which is comment #4 related to), but to be sure, Ronnie could you provide evm.log content with full entries related to instance provision which failed?
I have provided complete evm.log in attachments
Created attachment 1202848 [details] evm.log part related to instance provison with AZ
Now I see it, thanks Ronnie. Uncompress the bz, xz and tar archive evm.log.bz was a bit pain, but I finally succeed:) AZ name "volume-nova" occurs in your log too. This BZ looks to be duplicate bug of https://bugzilla.redhat.com/show_bug.cgi?id=1373559
Hm, interesting. Looking at the other BZ, it seems as if, Marek, you're selecting an AZ that doesn't exist; if you don't select it, provisioning works fine. Ronnie, is that the same case here? If so, the BZ becomes "make sure that phantom AZs are not created".
In case not selecting AZ the instance provisioning workflow works
*** Bug 1373559 has been marked as a duplicate of this bug. ***
https://github.com/ManageIQ/manageiq/pull/11416
https://github.com/ManageIQ/manageiq/issues/12155
verified on 5.8.0.4