Windows Server / SAN / Vertica Cluster

Navin_C

Hello All,

Might be  a generic question:

What can be the best storage layer/ staging layer for Vertica

I want to have a fast way of loading bulk data which can be done from COPY.

I want to understand what might be the best staging space for files to do this
The files will be exported from SQL SERVER

1. Windows Server(file exported to SQL server and from here we pick up files)
2.SAN box(file exported to SAN layer and then using the COPY from local command)
3. Vertica cluster nodes itself(file exported to one of Vertica nodes and then picked up by COPY command)

Thanks

[Deleted User]

If performance is your main concern, Vertica cluster nodes is probably best.  The down side there is the COPY statement is slightly more complex because you have to state which node should load which file.

I don't see the difference between SAN and Windows server from a Vertica perspective... they both just seem like network storage.  If that is where they reside, the file must cross the network at the time of the COPY statement.  Perhaps SAN is more robust/reliable though.

SAN is good because it simplifies the COPY statement, just tell Vertica that any node can load any file.  The downside is, as mentioned above, moving the file during the copy.