[
https://issues.apache.org/jira/browse/AMBARI-5589?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13982962#comment-13982962
]
Dmytro Shkvyra commented on AMBARI-5589:
----------------------------------------
STR:
1) Install 2 node cluster (minimum) with HDFS and YARN. Check http://<ambari
hosname>:8080/api/v1/clusters/<cluster
name>/services/YARN/components/NODEMANAGER. Request responce should have list
NM host_components and status of NM should be STARTED
2) In menu "Hosts">"All Hosts">"Node Manager" on Hosts perspective select
Decommision menu Item
3) On host, which has ResourceManager check /etc/hadoop/yarn.exclude file. In
this file should be present list of hosts where NodeManagers have been
decommited.
4) Go to Host perspective and select in menu of decommited NM "Delete" menu
item
5) Check http://<ambari hosname>:8080/api/v1/clusters/<cluster
name>/services/YARN/components/NODEMANAGER. There is host with deleted NM
should be dissapeared. Also check /etc/hadoop/yarn.exclude file. Hostname
where was deleted NM host component should be dissapeared also.
6) Repeat step 5 for all hosts where was installed NM hostcomponent
7) Check http://<ambari hosname>:8080/api/v1/clusters/<cluster
name>/services/YARN/components/NODEMANAGER. There is not should be any hosts in
host_components list and status of NM service componet should be INIT. Check
also /etc/hadoop/yarn.exclude on host with ResouceManager. This file should be
empty.
8) Add NM back on some hosts
9) Start NM's and run smoke tests
*Desired result:*
NM's were started successfuly and smoke test passed.
> Cannot start NodeManager after decommissioning, deleting, and then adding it
> back
> ---------------------------------------------------------------------------------
>
> Key: AMBARI-5589
> URL: https://issues.apache.org/jira/browse/AMBARI-5589
> Project: Ambari
> Issue Type: Bug
> Components: controller
> Affects Versions: 1.6.0
> Reporter: Dmytro Shkvyra
> Fix For: 1.6.1
>
>
> I decommissioned a NodeManager on a host, and then deleted it.
> Then I added NodeManager back to the same host.
> NodeManager will no longer start on this host, because from the
> ResourceManager's perspective, this host is still in the exclude list.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
