configuration file

[udayk_2007]

Hello Friends

Is it possible to use a different configuration file, for a particular stage in a job ?

For ex : The job is using a four node config file ,but a particular stage is needed to be executed on a 6 node configuration file. Is that possible ?

Thanks in advance.

Regards
Ulhas

[ray.wurlod]

That depends somewhat on your configuration file. To achieve what you've stated the configuration file would need to have (at least) six nodes but only four of those in the default node pool.

If your configuration file only has four nodes mentioned, there is no way that your requirement can be accomplished. Pools can reduce the number of nodes in which an operator executes, but can never specify more nodes than exist in the configuration file.

[jwiles]

I posted this reply into the wrong thread earlier today. It describes the methods to reduce/constrain the number of nodes a stage will use when a job is run, as mentioned in Ray's post.

You can constrain a stage to run on only certain nodes of a configuration file by either:

1) Defining a node pool in the configuration file and specifying that pool in the stage's Advanced tab
2) Specifying a "Node Map Constraint" on the Advanced tab.

Option 2 requires that you have supplied the name of the configuration file with the appropriate node names during the job design phase. I don't recall if option 1 has the same requirement.

The Parallel Job Developer Guide describes configuration files, node pools and using them in your job design.

Regards,

[ray.wurlod]

In my experience everything with pool names requires the configuration file to be named at design time. This is frustrating, because I prefer to use $APT_CONFIG_FILE with its default set to $PROJDEF, but I have to change this in order to effect the design, then change it back once the design is ready.

[chandra.shekhar@tcs.com]

Yes you can use the different node for your particular stage.

consider you have one configuration file having 10 nodes out of that 6 node you can use for particular stage using pools "DB2" and remaing 4 nodes for other stages(default).

In particular stage,click on stage icon --> advance setting--->select node pool & constraint ----> select DB.

and run Job.

go to director and check log that for perticular stage you wil get 6 connection.

Thanks
kishor