I have a large table (on the order of 1 TB) that I need to copy from one cluster to another. The cluster it currently resides on is 6.1.2-0. The target cluster is 7.0.0-1. The clusters are both on our LAN. They have different hardware and number of nodes (6 on the source, 4 on the target).
Just looking at the Admin Guide, I assume the best way to transfer the data is via a COPY or EXPORT command. My colleagues think it's possible a faster way is to somehow use backup/restore.
Is it possible to use backup/restore to copy table data from one cluster to a different one with a different number of nodes? I understand from the Admin Guide for 7.x that it's possible to backup a single object (here, table), but I haven't been able to take the backup/restore sections and conclude that there's a reasonably clean way to use backup/restore to move data from one cluster to another.
copy table from 6.x cluster to 7.x cluster
Moderator: NorbertKrupa
- JimKnicely
- Site Admin
- Posts: 1825
- Joined: Sat Jan 21, 2012 4:58 am
- Contact:
Re: copy table from 6.x cluster to 7.x cluster
I'd try the COPY FROM VERTICA command. Check it out in the online 7.0 SQL Guide (page 985)...
Here's an excerpt:
Here's an excerpt:
https://my.vertica.com/docs/7.0.x/PDF/H ... Manual.pdfYou can import data from an earlier HP Vertica release, as long as the earlier release is a version of the last major release. For instance, for Version 6.x, you can import data from any version of 5.x, but not from 4.x.
Jim Knicely
Note: I work for Vertica. My views, opinions, and thoughts expressed here do not represent those of my employer.
Note: I work for Vertica. My views, opinions, and thoughts expressed here do not represent those of my employer.
Re: copy table from 6.x cluster to 7.x cluster
Thanks for your reply.
Is there any particular reason I should favor going over the private vs public IP network? I assume the private network is used for the nodes to communicate with each other; if I go over the private network, will it slow things down on the source cluster by creating traffic on the private network?
Is there any particular reason I should favor going over the private vs public IP network? I assume the private network is used for the nodes to communicate with each other; if I go over the private network, will it slow things down on the source cluster by creating traffic on the private network?