Job Freeze..

swades · Post by **swades** » Mon Mar 12, 2007 12:43 pm

Hi All,
I am running server job and my job is freezing at certain point,
can somebody tell me why this happen,
i got this message in log file,

CopyOfJ_NTWX_FACT_ORDER_DETAILS..LC_MERGE_ROWS: DSD.StageRun Active stage starting, tracemode = 0.

Thnaks.

DSguru2B · Post by **DSguru2B** » Mon Mar 12, 2007 12:45 pm

What do you mean freezing? Is the designer client freezing or the job is hanging? What additional messages do you get in the log file? Reset the job and see if you have any messages "From previous run ..."
Need more info on design as well.

swades · Post by **swades** » Mon Mar 12, 2007 12:49 pm

I reset and run this job 3 times,
Even i did not get any warning or fatal errors,
still my job stops at this point.
i dont have any other information.
Thanks.

DSguru2B · Post by **DSguru2B** » Mon Mar 12, 2007 12:56 pm

Then how do you expect us to help you? No info of even the design, what your trying to do in the job, the data volume etc etc etc

swades · Post by **swades** » Mon Mar 12, 2007 1:06 pm

In out job 18 hash file we merge to a single hash file ,so i think problem with hash file reading or link collector stage is slow,
we collect through "Round Robin" type.
so help me Thanks.

DSguru2B · Post by **DSguru2B** » Mon Mar 12, 2007 1:09 pm

What happens when you write to the same hashed file from these 18 hashed files without the link collectory?

swades · Post by **swades** » Mon Mar 12, 2007 1:14 pm

Hi,
I did not try for that,
actually this job is design by another person,
and i cannt change in it,so here the design,

18 hash files----> link collector---->a single hash file.

my job ,every time freeze at this point and its shows blue link and there is not a single warning or fatal error.

so reply me thanks.

DSguru2B · Post by **DSguru2B** » Mon Mar 12, 2007 1:17 pm

Dont change it. Just build another job to do the same except, without the link collector. This is just for debugging purposes.
I have had issues with link collectors and hashed files (connected together) before and therefore dont like them next to each other.

swades · Post by **swades** » Mon Mar 12, 2007 1:48 pm

DSguru2B wrote:What happens when you write to the same hashed file from these 18 hashed files without the link collectory?

how could i do this job?

swades · Post by **swades** » Mon Mar 12, 2007 1:49 pm

swades wrote:
DSguru2B wrote:What happens when you write to the same hashed file from these 18 hashed files without the link collectory?
how could i do this job?

i mean with out collector stage how can i merge 18 files.

DSguru2B · Post by **DSguru2B** » Mon Mar 12, 2007 1:52 pm

Multiple writes to a hashed file is possible. Let your source hashed files go to either
- a single hashed file stage with the same hashed file name for each input
or
- 18 hashed file stages with the same name.

swades · Post by **swades** » Mon Mar 12, 2007 4:05 pm

DSguru,
but single Hash file reads "Round-Robin"(sequentially) or it just append all hash file,if it append all hash file then ideally it taking more time,because it wait for all file to read.
so answer me,actually my problem was solve.
Thanks.

DSguru2B · Post by **DSguru2B** » Mon Mar 12, 2007 4:53 pm

No my friend, it will do simultaneous reads from all of your source hashed files and load to a single hashed file.
How did you fix your problem? What was the issue?

kumar_s · Post by **kumar_s** » Mon Mar 12, 2007 10:09 pm

Hope your design is many to one. I mean many Hashed file to to single Hashed file. I also assume the output file is different from the source files.
That case, if any one source has large data to be transfered, the job takes that long.
And by the way, what gets your solution.
And you can mark the topic are resolved.

ray.wurlod · Post by **ray.wurlod** » Tue Mar 13, 2007 6:25 am

A hashed file is a way to implement a database table. You can have multiple streams inserting rows into it.

There is no need at all for a Link Collector stage.