Errors

DS_MJ · Post by **DS_MJ** » Tue Mar 29, 2005 6:45 pm

Hello:

Currently I am editing an existing job. That is adding a couple of extra columns.

The job flow consists of SourceSeqFile --> Link Partitioner--> Shared container-->Link Collector-->TargetSeqFile.

A hash lookup file going into the Shared Container.

Made the respective changes in the above stages. My job compiled successfully. When I run my job in the Director it shows that the job is aborted. In the staging area the link line from:

1. SourceSeqFile to LinkPartitioner is red
2. LinkCollector to TargetSeqFile is red.

Error in log file:Filename..Collect: |Error 11 in GCI Link initialisation.|

(NOTE: First_Col is one of the columns that I added.)

Advice would be appreciated.

Thanks in advance.

MJ

ray.wurlod · Post by **ray.wurlod** » Wed Mar 30, 2005 1:58 am

How many links run from the Link Partitioner stage into the Shared Container? How many links run from the Shared Container to the Link Collector stage? What algorithms are you using for the partitioning and collection? Is the Hashed File stage inside or outside the Shared Container?

DS_MJ · Post by **DS_MJ** » Wed Mar 30, 2005 9:10 am

Hi Ray:

How many links run from the Link Partitioner stage into the Shared Container? 4 links. There are 4 shared containers.

How many links run from the Shared Container to the Link Collector stage? One link going out of each shared container to the collector. 4 links coming into collector.

What algorithms are you using for the partitioning and collection? Round Robin.

Is the Hashed File stage inside or outside the Shared Container? Outside the shared container.

Is there a know bug with 7.5 version of Link Partitioner/Collector...?

Thanks,
MJ

DS_MJ · Post by **DS_MJ** » Wed Mar 30, 2005 5:31 pm

Hello:

Can Anybody help me to understand what the Link Error means.

Thanks in advance.

MJ

DS_MJ · Post by **DS_MJ** » Wed Mar 30, 2005 6:44 pm

Hello:

Read on some of the searches for the same kind of error. And I have set the Inter process buffer in the job.

Although, the meta data differs in the sense that source file has 2 extra columns. Only in the shared container's transformer do I not pass the 2 extra columns to output file. So in transformer input file has 20 columns and output column has 18 columns. These 18 columns go into the link collector and then the target file. Do you think this is the reason why I am getting the error.

Thanks in advance.
MJ

DS_MJ · Post by **DS_MJ** » Wed Apr 06, 2005 9:09 am

Hello All:

I think I have the solution to this problem. The issue was not link collector or version issue with 7.5 but the meta data and the linking from shared container to the link Partitioner/Collector. I deleted the shared containers, relinked and validated the links to other stages, verified that the meta data was same for all stages and now when I compile my job it does not give the error.

MJ

ray.wurlod · Post by **ray.wurlod** » Wed Apr 06, 2005 4:58 pm

Probably THE most important rule for Link Collector and Link Partitioner stages is identical metadata on all input and output links.

You have encountered one of the things that often traps folks new to shared containers. I make it a practice to document the columns being processed in a shared container, in the shared container's properties (and, often, in an Annotation on the shared container's canvas).