Bug #75192 Restore everything option for ndb_restore
Submitted: 12 Dec 2014 11:43
Reporter: Daniël van Eeden (OCA) Email Updates:
Status: Open Impact on me:
Category:MySQL Cluster: Cluster (NDB) storage engine Severity:S4 (Feature request)
Version:7.2 OS:Any
Assigned to: CPU Architecture:Any
Tags: ndb_restore, restore

[12 Dec 2014 11:43] Daniël van Eeden
For ndb_restore you need to restore the metadata once and the data multiple times (1x per datanode in the original cluster).

It would be nice if ndb_restore would detect the number of data nodes from the backup and then try to restore the metadata and data for all nodes.

Most backup solutions only need one command to do the restore, so people are used to that.

How to repeat:
Create a backup and then copy the backups from all nodes to the node where you're going to restore.

Now you need to run datanode_count+1 commands to do the restore.

Suggested fix:
If the backup from a datanode is missing report that:
ndb_restore: INFO: restored metadata
ndb_restore: INFO: restored data from node 1,2,3
ndb_restore: WARN: backup from nodeid 4 is missing, skipping.