Recommendation for the number of nodes

[hopewell]

Hi!

I am very new to Vertica and need to make a recommendation on the number of nodes we need in our database. Is there any where I can read about this? Or does any one know off the bat the number of nodes we'd need for 6 TB of data?

Thanks in advance!

[nnani]

Hello,

While performing the Capacity planning for vertica, you have to consider many parameters with respect to the Project.

1. The configuration of your servers ( RAM, Cores )
2. RAID level to be used
3. Total amount of RAW data
4. Growth percentage of data per year
5. Always remember it is recommended to use only 60 % of your disk space on your node to store data in Vertica. ( the rest will be used by Vertica for the internal operation )

Also the thumb rule, which I know, when dealing with Vertica Sizing is "The more complex the queries to be run on Vertica, The less the amount of data to be stored on each node "

Accordingly you have to calculate the per node data residing size for each node in Vertica and then plan on the number of nodes required by Vertica for 6 TB.

Hope this helps you.

[hopewell]

Thank you for getting back to me, nnani!

Great explanation!

[JimKnicely]

Hi,

I'd like to add that you may also need to keep space available on a server to store the files you'll be loading into the database via the COPY command (unless you are loading data from a remote server via the COPY LOCAL option).