Advanced Properties

[Nagaraj]

Hi There,

When do we use Node pool and Resource constraints? I just need one example where I can use and improve the Performance of my Jobs?

Thanks

[Nagaraj]

Sorry for the bold letters......! i didn't notice that i clicked the bold button.

[ray.wurlod]

If you have specified node or resource pools in your configuration file then you can constrain any stage to execute only in those pools. They are not a whole lot of value in an SMP (share everything) environment. However, an example: if you need to execute a stored procedure once (rather than once per node), you might constrain the Stored Procedure stage to execute in a single-node node pool.
In a multiple machine environment these come into their own. For example you might want to access a DB2 table partitioned by month, and only have DB2 client software installed on 24 machines. Your configuration file might therefore have those 24 machines in the DB2 node pool and more machines in the default node pool to look after processing.
Another example is that some machines may have more memory available than others; you might allocate these machines to the sort node pool so that sorting only occurs on them (and can use larger memory limits).

[Nagaraj]

Awesome ................Thanks Ray.....!

[throbinson]

Here's a Teradata example; If you know exactly how many rows will be returned from Teradata and this number of rows is less then the number of nodes, a warning will be generated because at least one node will not get any data. If you constrain by Node Pool to the number of rows you know will be returned then you will not get the warning. We pull single row resultsets from Teradata for Metadata (batch-id) and constrain on a single "Teradata" node pool when using a multi-node Configuration file.