Do I need to "lock" the JIRA instance when doing a native database backup?

[eacwpku]

It's almost required to lock down your JIRA instance when performing a full XML backup, so that the data isn't in an inconsistent (unrestorable) state. It's also recommended to use native database backups for disaster recovery. But, I don't see any recommendation about locking down your JIRA instance during native backups.

I don't know exactly how my native tool works, but what should I do to ensure a restorable native database backup?

[Nic Brough -Adaptavist-]

The easiest answer is "Stop Jira" - same as the XML export. Then you know no updates are going on

Some databases support "point in time" backups - that is you kick off a backup at say 10pm, and the process doesn't take any data that changes after 10pm, so you end up with a snapshot, exactly as you get with "stop Jira"

Another option is replication (quite common in large installations). You've got a second database server which takes data replicated from the first. When you want a backup, you break replication (so the master carries on being Jira and the slave has no updates going on), backup the slave database and then reenable replication at the end

There are even more clever options too, but I'd read up on your chosen database to find out what they are. Or better, take your local friendly DBA for a pint!