I've recently deployed several RHEL 7 servers on azure and set up backup jobs in the recovery service. As I understand the service, these backups are filesystem consistent. Unfortunately this is not enough and I need to make the backups application consistent.
Since azure backups are snapshot based, I assume this works similar to the way it does in VMware: By calling a script before and after taking the snapshot. The script is responsible to make the relevant applications flush and halt for the time it takes to make the snapshot.
In fact, I can see the freeze/thaw scripts that are called during the backup and I'm sure I could hack them to add this feature. Of couse, just because I could doesn't mean I should.
So the question remains: How do I make the backup jobs of my linux azure VMs application consistent?