Namenode HA Wizard fail and is not possible to roll back

[#1]

SYMPTOM: While performing step for Namenode HA, on the step of 'hdfs namenode -initializeSharedEdits' it failed with below error -

ERROR:

[root@localhost conf]# sudo su hdfs -l -c 'hdfs namenode -initializeSharedEdits'
16/11/22 09:43:56 INFO namenode.NameNode: STARTUP_MSG:
/************************************************************
STARTUP_MSG: Starting NameNode
STARTUP_MSG:   host = locahost/10.200.206.12
STARTUP_MSG:   args = [-initializeSharedEdits]
STARTUP_MSG:   version = 2.7.1.2.3.0.0-2557
STARTUP_MSG:   classpath = /usr/hdp/2.3.0.0-2557/hadoop/conf:/usr/hdp/2.3.0.0-2557/hadoop/lib/commons-cli-1.2.jar:/us           r/hdp/2.3.0.0-2557/hadoop/lib/xmlenc-0.52.jar:/usr/hdp/2.3.0.0-2557/hadoop/lib/jsch-0.1.42.jar:/usr/hdp/2.3.0.0-2557/  
.
.
.
STARTUP_MSG:   build = [email protected]:hortonworks/hadoop.git -r 9f17d40a0f2046d217b2bff90ad6e2fc7e41f5e1; compiled by            'jenkins' on 2015-07-14T13:08Z
STARTUP_MSG:   java = 1.8.0_51
************************************************************/
16/11/22 09:43:56 INFO namenode.NameNode: registered UNIX signal handlers for [TERM, HUP, INT]
16/11/22 09:43:56 INFO namenode.NameNode: createNameNode [-initializeSharedEdits]
16/11/22 09:43:56 ERROR namenode.NameNode: No shared edits directory configured for namespace null namenode null
16/11/22 09:43:56 INFO util.ExitUtil: Exiting with status 0
16/11/22 09:43:56 INFO namenode.NameNode: SHUTDOWN_MSG:
/************************************************************
SHUTDOWN_MSG: Shutting down NameNode at localhost/10.200.206.12
************************************************************/         

ROOT CAUSE: While performing NN Ha we found that Step 5 "Configure Components" is getting executed in a second, which was issue suspected. It was not stopping services and performing desired steps mentioned in screenshot below -

RESOLUTION: Suspected that Ambari server was holding cache in DB. Clearing the Ambari server cache using below steps resolved the issue -

# log out from the ambari server 
# list persisted k-v(ui cache) 

$curl -i -u admin:admin -H 'X-Requested-By: ambari' -X GET http://localhost:8080/api/v1/persist 

# get the cached state: 
$curl -i -u admin:admin -H 'X-Requested-By: ambari' -X GET http://localhost:8080/api/v1/persist/CLUSTER_CURRENT_STATUS 


# reset/clean the cache 
$curl -u admin:admin -H 'X-Requested-By:ambari' -d '{"CLUSTER_CURRENT_STATUS": "{\"clusterState\":\"DEFAULT\"}" }' -X POST 'http://localhost:8080/api/v1/persist'