You should consider monitoring the key processes for each service. This could be as simple as configuring your monitoring service to run ps --aux on the service at regular intervals to ensure that the process is alive. Another option is to use a monitoring library that supports process monitoring and the ability to automatically restart services if they crash. Deploying methods that allow monitoring and automatic recovery of the OpenStack processes will save you time and headaches as you troubleshoot.

To assist in disaster recovery, administrators will want to take a backup of the critical files for each OpenStack service, along with the database. For details on how to take a backup of the database, refer to the section earlier in this chapter. To take a backup of the critical files for each service, you will need to look into three particular directories for each service. The directories you want to consider include the following:

For example, to take a backup of the Glance image service, you would consider taking a backup of the files located at /etc/glance, /var/lib/glance and /var/log/glance.