Can you restore a schema backup to another cluster?
We have multi-tenants, each with a schema on one cluster. For disaster recovery I would like to backup each schema individually and restore each schema to an offsite cluster. I would also like to know if I can backup a schema on a 4 node cluster and restore the schema to an 8 node cluster? Mahalo, JackG
0
Comments
http://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/AdministratorsGuide/BackupRestore/CreatingObjectLevelBackups.htm?Highlight=object%20backup
Thanks for the quick reply. I had already read that doc but it does not answer my question.
Nor does this restore doc http://my.vertica.com/docs/7.0.x/HTML/Content/Authoring/AdministratorsGuide/BackupRestore/RestoringO...
Since it only refers to restoring to the *same* cluster that the schema backup was taken from ("To restore an object-level backup to the database from which it was taken"...).
I am asking 2 questions.
1. Can I restore a *schema* backup to a cluster that is not the cluster the database was taken from?
2. Does the secondary/restore cluster have to be the exact same number of nodes as the source/backup cluster?
Stated differently; If I backup a schema on clusterA that has 4 nodes. Can I restore the schema backup to ClusterB that has (more) 8 nodes?
3rd question for extra points ;-) Does the schema restore require the exact same database name as the backup cluster database name it came from?
Mahalo, - JackG
http://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/AdministratorsGuide/BackupRestore/RestoringObjectLevelBackups.htm?Highlight=restore
See the section - Restoring Objects to a Changed Cluster Topology
It would be nice if we had a copy schema equivalent to copy cluster (database).
Mahalo, - JackG
As Raul pointed out, we don't have this functionality currently in Vertica.
To answer your latest question, we can't bypass the need to same IPs as well. The the backup can oly be restored to the cluster where it was taken.
If you are looking to migrate data to another DB running on another cluster, I would suggest trying Vertica's Export/Import feature.
Administrator's Guide --> Copying and Exporting Data --> Exporting Data
https://my.vertica.com/docs/7.0.x/HTML/index.htm#Authoring/AdministratorsGuide/CopyExportData/ExportingData.htm?Highlight=Export
You will have to do this table wise, but it should fulfill your requirement.
Thanks,
Pravesh
Do have any further tips on increasing throughput using Exporting/Copying with 'CompressNetworkData'?
And/or how this performs relative to alternatives like "Parallel Load Streams"
Or loading data located on multiple nodes with a single copy statement?
Mahalo, - JackG
CompressNetworkData would be a good thing. However, please ensure that this parameter is set same on both the clusters (source and target).
We had a bug in Vertica v6.0 where target cluster would crash if we tried to import from other database and if the parameter was set differently on other cluster. This bug was fixed ib v6.1.
Export/Import throughput would always be lower than Parallel load streams as Export/Import implementation isn't parallel as yet.
Thanks.