Ok, these are my findings.
Backup is no problem using ssgbackup.sh script.
The trick with the restore is to take all gateway services down before restoring.
Restore on a cluster with two gateways;
1. Stop mysql slave on both hosts
2. Stop gateways on both hots (N.B services must be stopped using "service ssg stop", or else the database restore will fail on gateway 1)
3. Restore gateway 1 (db included in the backup image) with ssgrestore.sh
4. Restore gateway 2 (db not included in backup image) with ssgrestore.sh
5. Manually reboot both gateway hosts