Page 1 of 1

Multiple containter instances with standardize

Posted: Mon Feb 25, 2008 6:47 pm
by StephenB
We just tracked down some fatal errors to the QualityStage Standardize Stage when using multiple instances of a shared container. We have a container that validates an address, rejects invalid rows, replaces nulls with default values, etc... In a parallel job, we have 2 entities coming in on a row, so we split them using copy and sent each entity to the validating container and then to separate outputs. We saw the following error instantly:

Code: Select all

Unable to link rule file, VPHONE.UTO, file may need provisioning
( also saw with USADDR.UCL and USAREA.PAT as we removed stages to narrow down the problem )

and on the next run:

Code: Select all

From previous run 
DataStage Job 133 Phantom 12577
Job Aborted after Fatal Error logged.
Program "DSD.WriteLog": Line 305, Abort.
Attempting to Cleanup after ABORT raised in job STD_TEST_JOB.

DataStage Phantom Aborting with @ABORT.CODE = 1
I've read that shared containers are design time objects and are replaced when compiled, but they seem to cause a problem when the container is used multiple times in one job and contains the Standardize Stage. In testing we found there is not a problem using the Standardize stage twice in the same job or using a container multiple times in the same job without a Standardize stage.

Has anyone run into this before? Is this a bug/limitation or are we doing something wrong? Any ideas would be appreciated as we are planning to remove Standardize from our shared containers tomorrow.

Posted: Mon Feb 25, 2008 7:24 pm
by ray.wurlod
Provisioning is the step of moving your QualityStage jobs and stages (and rule sets, as necessary) from the Designer to the server.

Have you done this?

Posted: Thu Nov 06, 2014 1:03 am
by Chennai01
Can you please share the resolution as I am facing the similar situation.

Shared container which has Address standardize stage has been called multiple times with in the same parallel job and i am getting the error "Unable to link rule file, USADDR.SET, file may need provisioning".

Thanks

Posted: Thu Nov 06, 2014 3:49 pm
by ray.wurlod
... and HAVE you provisioned the Rule Set ?

Posted: Thu Nov 06, 2014 8:20 pm
by Chennai01
Yes, I did the provision all.

When i use one standardize stage in a job, it is working fine. But when i use two standardize stage in the same job, i am facing this issue.

"Unable to link rule file, USADDR.SET, file may need provisioning"

Below is my design.

DB(Input) --> Transformer --> Sharedcontainer(Address standardization) --> Copy stage --> Sharedcontainer(Address standardization) --> Transformer --> Dataset.

Sharedcontainer(Address standardization) has been designed earlier before an year and the input is 3 fields and output is 5 fields.

Copy stage is used to rename the columns which we receive from the first Sharedcontainer(Address standardization).

I am using 2 times as a single record has 2 different set of addresses and need to standardize. As such the shared container can handle only one set of address which made me to use 2 times and also we are using RCP for this job.

Posted: Thu Nov 06, 2014 9:31 pm
by Chennai01
Yes, I did the provision all.

When i use one standardize stage in a job, it is working fine. But when i use two standardize stage in the same job, i am facing this issue.

Below is my design.

DB(Input) --> Transformer --> Sharedcontainer(Address standardization) --> Copy stage --> Sharedcontainer(Address standardization) --> Transformer --> Dataset.

Sharedcontainer(Address standardization) has been designed earlier before an year and the input is 3 fields and output is 5 fields.

Copy stage is used to rename the columns which we receive from the first Sharedcontainer(Address standardization).

I am using 2 times as a single record has 2 different set of addresses and need to standardize. As such the shared container can handle only one set of address which made me to use 2 times and also we are using RCP for this job.

Posted: Fri Nov 07, 2014 1:03 am
by ray.wurlod
Seems there might be some kind of contention issue. What has your official support provider advised?