It is vital that you backup your system frequently so that you can recover the system in the event of administrator error, catastrophic hardware failure, fire, etc. Disk mirroring on its own is not sufficient as it does not protect against most failures (it only protects against a single disk failure). The following backups are recommended:
- Copy /var/lib/enswitch to a remote system daily. If using a cluster with multiple machines, this should be done from the NFS server. This will copy:
- Backups of the database, as Enswitch writes its own database backups to /var/lib/enswitch/backups.
- All other customer data files such as invoices, voicemails and greetings, call recordings, IVR sound files, etc.
- Backup each machine either weekly or monthly, and also after any major upgrades. This allows you to recover any machines that are completely lost. At minimum you must back up /etc, but an entire backup of all local filesystems is recommended.
The actual method used to do the backups is up to the system owner. Using rsync to an off-site machine is both safe and convenient, but other methods such as tapes or shared disks are also feasible.