ansible start process if not running using ps - linux

Question

i am using ansible to start the namenode process if incase its not running, I am using jps to check for the process and check the status using service_namenode_status and start the process using service_namenode_start any idea how to use this properly...?
- name: Check if Namenode is running
shell: jps | grep " NameNode" | grep -v grep
ignore_errors: yes
changed_when: false
register: service_namenode_status
- name: Report status of Namenode
fail:
msg: |
Service NameNode is not running.
Return code from `jps | grep " NameNode" | grep -v grep`:
{{ service_namenode_status.rc }}
when: service_namenode_status.rc != 0
register: service_namenode_start
- name: Start Namenode
command: su - hdfs -c "/usr/hdp/current/hadoop-client/sbin/hadoop-daemon.sh --config $HADOOP_CONF_DIR start namenode"
tags: namenode
when: service_namenode_status.rc != 0
OUTPUT:
TASK [namenode : Check if Namenode is running] *********************************
fatal: [server1.hdp.vagrant.data.uc]: FAILED! => {"changed": false, "cmd": "jps | grep \" NameNode\" | grep -v grep", "delta": "0:00:00.192755", "end": "2016-11-09 10:59:10.360552", "failed": true, "rc": 1, "start": "2016-11-09 10:59:10.167797", "stderr": "", "stdout": "", "stdout_lines": [], "warnings": []}
...ignoring
TASK [namenode : Report status of Namenode] ************************************
fatal: [server1.hdp.vagrant.data.uc]: FAILED! => {"changed": false, "failed": true, "msg": "Service NameNode is not running.\nReturn code from `jps | grep \" NameNode\" | grep -v grep`:\n1 \n"}
NO MORE HOSTS LEFT *************************************************************
RUNNING HANDLER [metastore : restart postgresql] *******************************
to retry, use: --limit #/Users/krisdigitx/myLab/hdp/provisioning/site.retry
PLAY RECAP *********************************************************************
server1.hdp.vagrant.data.uc : ok=86 changed=32 unreachable=0 failed=1
Ansible failed to complete successfully. Any error output should be
visible above. Please fix these errors and try again.

Answer 1

Try this:
- name: Check if Namenode is running
tags: namenode
shell: jps | grep " NameNode" | grep -v grep
ignore_errors: yes
changed_when: false
register: service_namenode_status
- name: Start Namenode
tags: namenode
command: su - hdfs -c "/usr/hdp/current/hadoop-client/sbin/hadoop-daemon.sh --config $HADOOP_CONF_DIR start namenode"
when: service_namenode_status.rc == 1
My changes were:
I added the "tags: namenode" line to the 1st part; you'll need to collect the register value if you are filtering on this tag
I removed your 2nd part and went straight to the 3rd.
A similar config is working for me for a similar problem.