Does DC + DR architecture with GTID enables reverse - replication?

[뤀쪼]

Hello, my team is currently using Galera clustesr for the Data Center segment and the Disaster Recovery segment. Normally, a node in DR segment replicates the data from a node in DC segment and I guess I can call it MariaDB replication that it is a node-node replication using binlog. The operation guy said he has never adopted any GTID option for our clusters.

The problem is when the DC encounters a problem and stops, then we have to use the DR cluster for a while. Later we find out that DR has more updated data so we need to replicate the data of DR cluster into DC cluster somehow.

The only way we have used is copying or dumping that takes couple of hours. My manager asked me to find out if there is a way to make DR node -> DC node replication smoother and faster and what I found was the global transaction identifier, so called GTID.

So I have two questions to ask,

1) Can GTID helps our DC node (that has stopped for a while) to replicate data from a DR node without sync-error?

2) if so, does MariaDB Galera cluster also provides the auto-positioning functions because I found out it is more limited for MariaDB Galera cluster than for MySQL Galera cluster.

Sorry for my English in advance, if there is any misconception.

Regards

Luke