How to use backup and restore feature for moving one database from one clsuter to another cluster

Help
1

Hi Team.
We are using memsql-ops 6.7 and no NFS mount for backup. We put it on local path. Since backup commanddoes partition backups on leabves as well. instead of using mysqldump or pipelines or export data feature, can I backup and restore methrd to move database from one cluster to another cluster. Lets say from Dev env to Test env clusters

2

The easiest thing is to have some shared storage (S3, Azure blob store, NFS) and backup to that and restore from that.

But if that is totally impossible then you can go copy the backup content from the backup folders on each leaf node on the source cluster, copy it to the target cluster, and restore it. This will require logging in to every leaf on the source and target.

3

Hello Hanson.
Can you please elaborate on this and provide me the steps to do it. Assume, I have copied all the backup contents of all leafs (4 leafs) and Agg (2 agg) on source cluster to target cluster (4 leaf and 2 Agg), now tell me what is the steps to restore those files to this new cluster.

“But if that is totally impossible then you can go copy the backup content from the backup folders on each leaf node on the source cluster, copy it to the target cluster, and restore it. This will require logging in to every leaf on the source and target.”

4

This is not documented or recommended, and it is a bit involved. I don’t have time right now to develop a complete walk-through example. But you can try something like this:

backup database foo to "/tmp";

then go to /tmp on each node at the Linux prompt and do:

ls -tla foo*

copy the files listed to the same location (/tmp) on the equivalent nodes on the target cluster. Use chmod to make them readable to all users or at least the ‘memsql’ user. Then do something like:

restore database foo_restore from "/tmp/foo.backup";