Internal data error

eoyylo · Post by **eoyylo** » Fri Jun 13, 2003 2:09 am

Hi,
i have a job that uses some hash files. The hash file's load process is performed by a different job. The job that creates the hash file ends with this message:

File '/tmp/EBIS/F_NORMALISED_EVENT_HSH_Adc/DATA.30':
Computed blink of 0x974 does not match expected blink of 0x0!
Detected within group starting at address 0x874FD800!

In the job i set the following parameters:
1)create file
2)allow stage create write cache
3)clear file before writing

and in OPTIONS

4)Delete file before create

The OS is Sun 5.8.

thanks in advance

Mario Loreti

WoMaWil · Post by **WoMaWil** » Fri Jun 13, 2003 6:49 am

Ciao Mario,

I don't have an ad-hoc answer for you:

What I would do is:

- to try with a shorter file_name
- how many links write to the file, if more than one, do make a dummystage at the beginning of the job for re-create or/and clear file.

What about your hard disk, is enough space left?

Wolfgang Huerter
=====================
Cologne, Germany

kduke · Post by **kduke** » Fri Jun 13, 2003 8:41 am

Mario

You are corrupting your hash file some way. Turn write cache off and try it. I was told that write cache does not work on certain versions of UNIX and they do not know why.

Thanks Kim.

Kim Duke
DwNav - ETL Navigator
www.Duke-Consulting.com

ray.wurlod · Post by **ray.wurlod** » Fri Jun 13, 2003 9:17 pm

The word "blink" in the error message is short for "backward link". When there is a blink error, the internal pointers in the hashed file have become corrupted in some way.
Kim's advice probably gives you the best workaround. I would also recommend logging a call with your support provider, who will be able to determine whether (or when) a fix is available on your version of DataStage on your particular UNIX (Solaris).

Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518