Question about memsql's DR

Off-Topic
1

for oracle’s DR (disaster recovery)
I heard that secondary cluster loads primary cluster’s PB-sized data from disk during initial setup.
and then build an ADG(ActiveDataGuard).
The data that is updated later is replicated over the network.

for memsql’s DR
Before COMMAND below
REPLICATE DATABASE db_name FROM root:root_password@primary-MA:3306;
I tried to load PB-sized data using either ‘NFS PIPELINE’ or ‘LOAD DATA IN FILE’ but it was not the right method.
Is there any other suggestion or way to load the initial massive amount of data from disk rather than from the network?

Is there any other suggestion or way to load the initial copy from disk rather than from the network?

2

Do you mean your load to the “primary” cluster failed to work before you even started replication?

LOAD DATA and PIPELINES can both be made to run at high scale. What do you mean it was not the right method? What errors happened?

3

We want to create a secondary cluster to build a disaster recovery .

I don’t want to replicate over the network of the command ‘replicate database’ because the primary cluster has a PB size.

Because replication over the network is slow, as in the case of Oracle above, the initial replica tries to use a different method.

We want to replicate a large volume more quickly and then perform the ‘replicate database’ command.

Are there any strategies involved?