Cloudera-scm-agent dead but pid file exist

Issue : cloudera-scm-agent failed to restart, saying agent dead but pid file exist. Sometimes it starts good but it`ll die after few min.

Solution : Try 1

Delete the pid file and restart the agent.

# rm -f /var/run/cloudera-scm-agent.pid

# service cloudera-scm-agent restart

# service cloudera-scm-agent status

It`ll work for most of the cases. If it doesn`t work try the next step.

Try 2 :

Delete the pid file and restart the agent.

# rm -f /var/run/cloudera-scm-agent.pid

Hard stop the agent :

# service cloudera-scm-agent hard_stop_confirmed

Hard restart the agent :

# service cloudera-scm-agent hard_restart_confirmed

Try 3 :

Check for zombie agent process.

Check the agent status :

# service cloudera-scm-agent status

If the agent status is stopped, there should not be any agent process running in background.

Check it using PS command :

# ps -ef | grep "cloudera-scm-agent"

Kill the zombie agent process

kill -9 <pid>

Start the agent :

# service cloudera-scm-agent start