Sandor Magyari created AMBARI-13712:
---------------------------------------
Summary: Blueprints: Support for Blueprint export of a running
Kerberized Cluster
Key: AMBARI-13712
URL: https://issues.apache.org/jira/browse/AMBARI-13712
Project: Ambari
Issue Type: Bug
Components: ambari-server
Reporter: Sandor Magyari
This task involves implementing the support for exporting a Blueprint from a
Kerberized cluster, and having this document contain everything necessary to
re-create this Kerberized cluster with new hardware.
An exported Blueprint for a Kerberized cluster will need to contain:
The existing component and configuration for the cluster (this can remain as it
currently is, no changes required)
The “kerberos-env” and “k2b5-conf” configuration types. This will largely occur
automatically, since these types are just plain configuration types on the
cluster, and will be exported along with the other config types. Note: There
will be some work required here in order to make sure that hostnames and
passwords are not included in the exported Blueprint. The
BlueprintConfigurationProcessor will need to be updated to properly mask out
the hostnames/passwords. This is similar to existing config type handling, and
will be a relatively small amount of work.
The Kerberos state of the cluster. The Blueprint processor will need to query
the “KerberosHelper” API to determine if a cluster is Kerberized. If Kerberos
is enabled, then the Blueprint processor will use the same API to obtain the
Kerberos descriptor associated with this cluster (found as a REST resource
under /api/v1/clusters/cluster_name/artifacts/kerberos_descriptor). This
Kerberos descriptor should be added to the exported Blueprint, as an embedded
descriptor inside the Blueprint .json file.
We should be able to export a Blueprint from a Kerberized cluster regardless of
the initial deployment type (Ambari UI used to create cluster vs. Blueprint
deployment used to create cluster).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)