Cleanup Alternatives

Cloudera Manager’s agent relies on the Linux alternatives framework to register service binaries and configuration paths. When that framework gets into a bad state, the agent can fail to update or start services, and the usual fix is to purge the stale alternatives entries and let the agent rebuild them from scratch.

Symptoms

You may notice the agent stuck in a restart loop, services failing to start after a Cloudera Manager action, or log entries referencing missing or conflicting alternative links under /etc/alternatives/.

Procedure

1. Identify the affected alternatives. List the entries that point into the Cloudera install tree:

   ls -l /etc/alternatives | grep "/opt/cloudera"

2. Stop the agent so it does not fight the cleanup:

   service cloudera-scm-agent stop

3. Remove the stale alternatives. This loop checks for a matching record in /var/lib/alternatives/, removes it if present, and then deletes the symlink in /etc/alternatives/:

   ls -l /etc/alternatives/ | grep "/opt/cloudera" | awk '{print $9}' | while read m; do
     if [[ -e /var/lib/alternatives/${m} ]]; then
       echo "Removing ${m}"
       rm -fv /var/lib/alternatives/${m}
     fi
     rm -fv /etc/alternatives/${m}
   done

4. Restart the agent. On the next cycle the agent will re-register all of its service alternatives cleanly:

   service cloudera-scm-agent start

Verifying the Fix

After the agent restarts, confirm the alternatives are back in place:

ls -l /etc/alternatives | grep "/opt/cloudera"

You should see the expected symlinks (for example cloudera-scm-agent, cloudera-scm-agent.service, and any service-specific entries) pointing into /opt/cloudera/. If the list looks correct and the agent reports a healthy status in the Cloudera Manager UI, the issue is resolved.

• Technology
• How-To

• SysAdmin
• Troubleshooting