Skip to main content
Version: 24.04

Back up and restore your pollers

Case n°1: Making backups to be able to install a new poller from scratch​

What you need to back up on your active poller​

In this list, we assume you have enabled the automatic installation of the Centreon plugins.

You need to back up the following elements to be able to rebuild your poller from scratch in case of a problem.

  • Custom plugins (example: community plugins, or custom developments)
  • If you are using the centreon-as400 connector (installation and configuration): backup /etc/centreon-as400/
  • Synchronize the following files regularly (e.g. using the rsync command):
    • /var/log/centreon-engine/retention.dat (up to every 15 minutes) to keep acknowledgements, downtimes and statuses.
    • /var/lib/centreon/centplugins/* (up to every 5 minutes) to keep the plugins cache.
    • /etc/centreon-gorgone/config.d/* (once) to keep the connection information to the central server.
    • /var/lib/centreon-gorgone/.keys/* (once) to keep the same fingerprint for ZeroMQ authentication.
    • /etc/snmp/centreon_traps/ (once a week), if you are using SNMP traps on the poller.
    • /etc/centreon/ (once a week).
    • /var/log/centreon-engine/ (once a day), as it contains retention.dat and all archives (necessary to rebuild the dashboards).

Switching to the new poller​

If your poller has died, install a new poller, in the correct version for your platform, attach it to the central server, then restore all the elements listed above.

If you didn't backup /var/lib/centreon-gorgone/.keys/*, you will need to change the poller's fingerprint in the configuration of the central server. Read the following article on our community platform The Watch.

Case n°2: "Standby" poller​

Another way to prepare for disaster is to have a "standby" poller on which you synchronize the data from your active poller. If your poller dies, just switch to the standby poller.

What you need to synchronize​

Install and configure the same elements on your standby poller as on your active poller, and synchronize your data as described here.

How to switch to the standby poller​

  1. Start the standby poller if it is stopped.
  2. Change the new poller's network configuration to give it the old poller's IP address.
  3. Restart gorgoned on the poller first, then on the central. 
    systemctl restart gorgoned
  4. Make sure there are no communication errors in Gorgone's log. 
    tail -F /var/log/centreon-gorgone/gorgoned.log | grep ERROR
  5. Deploy the configuration for the poller using the Restart method (in the Restart Monitoring Engine list).
  6. If you didn't backup /var/lib/centreon-gorgone/.keys/*, you will need to change the poller's fingerprint in the cache of the central server. Read the following article on our community platform The Watch.

Case n°3: VM snapshots​

If your poller is running on a VM, take regular snapshots of your poller. In case a problem occurs, restore the snapshot as with any virtual machine.