When I attempt to scale up a cluster that uses the CDH 5.3.0 plugin, it fails during the 'Update configs' step. I've included the failed event below:
{
"successful": null,
"tenant_id": "<tenant_id>",
"created_at": "2015-09-01T00:22:37",
"step_type": "Plugin: scale cluster",
"updated_at": null,
"cluster_id": "<cluster_id>",
"step_name": "Update configs",
"total": 1,
"id": "<id>"
},
From my research into this, I found that the section linked here https://github.com/openstack/sahara/blob/stable/kilo/sahara/plugins/cdh/cloudera_utils.py#L139-L155 is where the failure is occurring. This piece of code is using a thread pool to invoke the cloudera manager to deploy the client config. The thread pool is re-wrapping the exception so the real stacktrace never makes it into the sahara logs. The reason code I find in the logs is '__exit__'.
Fix proposed to branch: master
Review: https:/