DSXchange

I am using the Iway Enterprise stage to execute an Information Builders ReportCaster Job. The stage works fine, but it executes the job twice. Has anyone seen this behavior and if so, were you able to correct the problem?

Thank you,

Looking through the log, I see that my Parallel job's "main_program:APT Configuration file:" is set to C:/IBM/InformationServer/Server/Configurations/default.apt. Which looks like this in the log detail.

main_program: APT configuration file: C:/IBM/InformationServer/Server/Configurations/default.apt
{
node "node1"
{
fastname "0CNW1529"
pools ""
resource disk "D:\Pool\DataSets" {pools ""}
resource scratchdisk "D:\Pool\Scratch" {pools ""}
}
node "node2"
{
fastname "0CNW1529"
pools ""
resource disk "D:\Pool\DataSets" {pools ""}
resource scratchdisk "D:\Pool\Scratch" {pools ""}
}
}

I'm wondering if because there are two nodes, this stage is running the job twice. I created a new configuration file (apt file) that only has one node. Can someone tell me where to set the configuration file (apt) for a the job?

It's logically executing the job twice - the same job design executes on both nodes. However, partitioning makes sure that some of the records are processed on one node and the remainder are processed on the other node.

...it could be like MQ though...it might simply repeat the SQL statement and get duplicate rows..... check it carefully. ...and look in your Job Parameter env variables for the APT_CONFIG_FILE .....

Ernie

webfocusguy wrote:Can someone tell me where to set the configuration file (apt) for a the job?

As noted, you'd need to have added the $APT_CONFIG_FILE variable to the job, that would give you the ability to override it and have the job point to the single node config rather than the default two node config.

There are several that are considered a "Best Practice" to add to every job so you have the ability to override any of them as needed, temporarily or otherwise. This is one of them.

Changed the APT Config file and set the var but still runs on both nodes. Hmmm. I wonder why this stage is not available in a server job? I found out that someone on our team opened a case on this yesterday afternoon. Thank you everyone for the comments and for the tip on the var and I'll post the resolution when we get it.

Hopefully you created a new single-node config file rather than changing the default one... or at least did something to ensure you have two when you are done. If you'd like us to look at the new config, feel free to post it here. If things are still running on two nodes, then either the new config is suspect or you're still picking up the default one somehow.

Good question but I did create a copy of the config file. Below are the two to compare.

Other activities: In the Stage Tab of the IWay stage, under the "Advanced sub-tab", there is a drop-down box, "Execution mode:". It's set to Default(Sequential) but it's disabled and I can't change the value. I added the variable &APT_EXECUTION_MODE, thinking that might enable the drop-down but it didn't.

From the Job Properties Dialog, under the Parameters Tab, when you click on the Default Value field of the $APT_EXECUTION_MODE variable, you get the $AP_EXECUTION_MODE - Setup List and Default dialog box. At the bottom you can set the default value. I changed it from "Parallel" to "One Process" but it had no effect.

Curious thing about the APT EXECUTION MODE variable set up dialog box, the Add and Remove options are disabled. Starting to seem like a conspiracy to me.

=============================
main_program: APT configuration file: C:/IBM/InformationServer/Server/Configurations/default.apt
{
node "node1"
{
fastname "0CNW1529"
pools ""
resource disk "D:\Pool\DataSets" {pools ""}
resource scratchdisk "D:\Pool\Scratch" {pools ""}
}
node "node2"
{
fastname "0CNW1529"
pools ""
resource disk "D:\Pool\DataSets" {pools ""}
resource scratchdisk "D:\Pool\Scratch" {pools ""}
}
}
================================
main_program: APT configuration file: C:/IBM/InformationServer/Server/Configurations/singlenode.apt
{
node "node1"
{
fastname "0CNW1529"
pools ""
resource disk "D:\Pool\DataSets" {pools ""}
resource scratchdisk "D:\Pool\Scratch" {pools ""}
}
}

IBM Tech Support came through with this answer;

1) By default the iway operator will execute the open command twice.
2) This is to facilitate certain operations for setting iway server before executing the query. Please set 'Persistent' option that is available with Open command to true. This will result in the command being executed only once.

I tested this and it works as expected.