Errors in Cue should be propagated to the user

[Steve Leon]

When a request is issued (ie to create a cluster), if the error happens in cue, the cluster goes to ERROR but the user does not know the cause of it. At least, the user should be returned a message, or an ID (that the user can use to track the error).

This was reproed when a cluster was created and nova failed to create the VM for cue due to not having enough IP available. Here is the error from cue-worker.

2015-10-21 18:21:10.446 32477 WARNING taskflow.conductors.backends.impl_blocking [req-7476934f-08ee-4dbd-91a0-282c11d09d4b e3a07bc5275049ce8b8cc08e9b896c29 6b226ce3faba4d4992c5f96e9b198b92 - - -] 1. Failure: neutronclient.common.exceptions.IpAddressGenerationFailureClient: No more IP addresses available on network 35a29c74-7187-4ef4-b9fb-c0ac5e454c72.
Traceback (most recent call last):
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/taskflow/engines/action_engine/executor.py", line 67, in _execute_task
    result = task.execute(**arguments)
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/os_tasklib/neutron/create_port.py", line 59, in execute
    port = self.os_client.create_port(body=body_value)
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/neutronclient/v2_0/client.py", line 102, in with_params
    ret = self.function(instance, *args, **kwargs)
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/neutronclient/v2_0/client.py", line 544, in create_port
    return self.post(self.ports_path, body=body)
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/neutronclient/v2_0/client.py", line 298, in post
    headers=headers, params=params)
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/neutronclient/v2_0/client.py", line 211, in do_request
    self._handle_fault_response(status_code, replybody)
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/neutronclient/v2_0/client.py", line 185, in _handle_fault_response
    exception_handler_v20(status_code, des_error_body)
  File "/opt/stack/venv/cue-20151020T024829Z/lib/python2.7/site-packages/neutronclient/v2_0/client.py", line 70, in exception_handler_v20
    status_code=status_code)

The cluster went to ERROR but the user didnt know the cause. I had to look into the logs to find the issue.

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/244308
Committed: https://git.openstack.org/cgit/openstack/cue/commit/?id=1be42880706dd45d15686bc3eabeb58af546bfb0
Submitter: Jenkins
Branch: master

commit 1be42880706dd45d15686bc3eabeb58af546bfb0
Author: dagnello <email address hidden>
Date: Wed Nov 11 12:50:10 2015 -0800

    Adding support for error details on cluster creation

    Cue's current implementation does not provide any additional information
    to the user related to why a cluster went into ERROR state during
    cluster creation. This patch adds error_detail field in the
    clusters table and exposes this field through the API. The
    update_cluster_task now extracts failure flow details and forwards this
    information on cluster record update.

    Partial-Bug: 1508730
    (updates also required to Cue client and dashboard)
    Closes-bug: 1516735
    Change-Id: I440fec30bc3d051189585479b8d2c1219a98cf40