what is vertica projections and segments.

renumetukuru

I would like to understand How vertica achieves high availability and performance with those? how vertica segments data while loading and querying..?

[Deleted User]

Hi, These are some pretty fundamental questions, more than I think I could properly explain with just a forum post... I would encourage you to read through our online documentation: http://www.vertica.com/documentation/documentation-6-1/ Also, our training offerings (both free/online and in-person): http://www.vertica.com/customer-experience/12573-2/ http://www.vertica.com/customer-experience/training/ I think the free online "Technology Overview" training in particular, in combination with the documentation/etc, may help to answer your questions. Adam

renumetukuru

HI Adam, I have a 16 node cluster and having a table with 500k rows. When i check gte_projections('table0, am seeing 2 projections since k-safety is 1. also it shows segmented yes. How can i see, which cluster nodes having segemneted rows.. I mean among 16 nodes, how can i see on which nodes these 500k rows segmented?

Navin_C

Hello Renumetukuru, First of all, When you say a table is segmented on a cluster, that means the data in table is segmented across all nodes on the cluster. Vertica provides you option while segmenting data Segmented on All Nodes Automatically distributes data evenly across all nodes at the time the CREATE PROJECTION statement is run. The ordering of the nodes is fixed. Segmented on Offset An offset value indicating the node on which to start the segmentation distribution. Omitting the OFFSET clause is equivalent to OFFSET 0. The offset is an integer value, relative to 0, based on all available nodes when using the ALL NODES parameter. See example below. Segmented on NODES node [ ,... ] Specifies a subset of the nodes in the cluster over which to distribute the data. You can use a specific node only once in any projection. Now that you want to see on which nodes the 500 k rows segmented Try this query select projection_name, anchor_table_name, is_segmented, node_name from projections where anchor_table_name='table_name'; This will show you the projection for specific table and their distribution on nodes Hope this helps..