DSXchange

Hi All,
I have an OLEDB Stage with a select, on SQL Server, from 2 tables with a join (totaling over 140,000,000 rows).
a where part reduces the row count retrieved to about 18,500,000 rows (5 columns).
the result set is written to a seq file.

At first When I Had run this job it finished fine (twice I think)!!! and I got about 18,500,000 rows.

but ever since it aborts with "Abnormal termination of stage ChrInit..T_chr_998 detected"
after about 16,500,000 rows written to the seq file.

I tried deleting the seq file and rerunning the job but getting same Abort.

checked with system personal no disk space problems.

Any Ideas come to mind ???????

Thanks in advance,


Roy R.

Only some "non-logical" ideas, which have had for me already good expierences:

(1) Export job, delete Job, reimport job
(2) delete the error-making stage and rebuild it
(3) parallel the output for to write in 10 files (take the mod of a number field for to devide) and join the 10 files into one in after-job-routine.


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

Roy

You could be running out of temp space. If you are on UNIX then you could be hitting the file size limit of ulimit. What OS are you on and what version of DS? I would definitely break this up into smaller files if possible like Wolfgang stated. Run these in parallel would speed things up as well.

What is your UVTEMP folder? When it bombs then do a df command and see if this filesystem is not full.

Kim.