Troubleshooting HDF upgrade

Learn about issues that might occur during the upgrade process, and how to resolve them.

Issue: Host health error
The host health shows the following error:
Clock Offset Suppress...The host's NTP service could not be located or did not respond to a request for the clock offset.
Enable and start the NTP service for all hosts:
yum install -y ntp
systemctl start ntpd.service
systemctl enable ntpd.service
systemctl status ntpd.service
Otherwise perform the following steps:
  1. Click on the hosts health error, and navigate to the Host Clock Offset Threshold configuration.
  2. Configure the values for thresholds to Never.

  3. Click Save Changes.
Issue: Swappiness
The host check process checks how many pages were swapped in a given time, and an error might appear if that number is not reached, as shown in the following image:

Set swappiness to one for all hosts.
cat /proc/sys/vm/swappiness
echo 1 > /proc/sys/vm/swappiness
cat /proc/sys/vm/swappiness
It suppresses the swappiness alerts for the cluster. Additionally, you can perform the following checks:
  • Check configuration warnings for each service.
  • Review JVM parameters and configuration for all services (some services are not transitioned).
  • Review the Log4J configuration such as logs dir, size, and backup index.
Issue: Cloudera agent security
You might experience an issue with Cloudera agent security and Cloudera agents might not connect to the server.
You can perform the following checks:
  • Fix ownership of /var/lib/cloudera-scm-agent/agent-cert
  • chown cloudera-scm:cloudera-scm /var/lib/cloudera-scm-agent/agent-cert
  • chmod 755 /var/lib/cloudera-scm-agent/agent-cert