Partial online backup

Moderator: NorbertKrupa

heatlamp
Newbie
Newbie
Posts: 9
Joined: Mon Aug 20, 2012 11:59 pm

Partial online backup

Post by heatlamp » Tue Aug 21, 2012 12:19 am

I’m new to Vertica, but not to RDBMs.

For whatever reason, our 20TB Vertica database frequently falls over, and management would like to create a backup environment that only includes data specific to this division (less than 5% of the complete database). The backup system would only be used when the production system is down, and would not be collocated with production.

The question is how best to do this. In a traditional RDBMS, something like materialized views over a database link might be used.

The backup environment need not be Vertica, but it could be if that’s the most efficient solution. A primary concern is the performance impact on the production environment, which is often stretched.

Any options appreciated.

Thx, Jim

scutter
Master
Master
Posts: 302
Joined: Tue Aug 07, 2012 2:15 am

Re: Partial online backup

Post by scutter » Tue Aug 21, 2012 10:04 pm

HI Jim,

Have you looked into the root cause for why your cluster is falling over? Seems like you might be better off resolving that rather than creating a mirror of a subset of the Vertica data.

--Sharon
Sharon Cutter
Vertica Consultant, Zazz Technologies LLC

heatlamp
Newbie
Newbie
Posts: 9
Joined: Mon Aug 20, 2012 11:59 pm

Re: Partial online backup

Post by heatlamp » Tue Aug 21, 2012 11:32 pm

Thanks for the reply. I have no control over the production environment, but other people are looking into it.

My mandate is to create a backup solution.

Jim

User avatar
JimKnicely
Site Admin
Site Admin
Posts: 1825
Joined: Sat Jan 21, 2012 4:58 am
Contact:

Re: Partial online backup

Post by JimKnicely » Wed Aug 22, 2012 12:58 am

Hello heatlamp,

How many tables are involved? Have you looked into how to EXPORT data via a SELECT statement (see the "Moving Data Between Vertica Databases" section of the Admin. Guide). You could export the 5% of data off to another Vertica database on a regular basis. You can add a SELECT AT EPOCH LATEST expression to include data from the latest committed DML transaction. Although keep in mind that they'll be a lag in the data in PROD and the data available in your backup database since you may only move the data once a day.
Jim Knicely

Image

Note: I work for Vertica. My views, opinions, and thoughts expressed here do not represent those of my employer.

User avatar
JimKnicely
Site Admin
Site Admin
Posts: 1825
Joined: Sat Jan 21, 2012 4:58 am
Contact:

Re: Partial online backup

Post by JimKnicely » Wed Aug 22, 2012 1:02 am

heatlamp,

Out of curiosity, what version of Vertica is running in your PROD environment? And how many nodes are in the cluster? I'm with scutter in that we'd all like to help you figure out why the cluster maybe failing. And what no you mean by failing? Are nodes leaving the cluster? Sorry about all the questions...
Jim Knicely

Image

Note: I work for Vertica. My views, opinions, and thoughts expressed here do not represent those of my employer.

heatlamp
Newbie
Newbie
Posts: 9
Joined: Mon Aug 20, 2012 11:59 pm

Re: Partial online backup

Post by heatlamp » Wed Aug 22, 2012 1:04 am

Thanks again, but I really don't want to get off topic. Perhaps I or another will post RE our production issues.

Jim

User avatar
JimKnicely
Site Admin
Site Admin
Posts: 1825
Joined: Sat Jan 21, 2012 4:58 am
Contact:

Re: Partial online backup

Post by JimKnicely » Wed Aug 22, 2012 1:15 am

Ok, but what do you think about the EXPORT idea?
Jim Knicely

Image

Note: I work for Vertica. My views, opinions, and thoughts expressed here do not represent those of my employer.

Post Reply

Return to “New to Vertica Database Administration”