DSXchange

Go to the TCL prompt by logging on to the administrator and run CLEAR.FILE RT_LOGnnn This might not be the best way as it will remove the default purge settings. Once you are able to take care of all the huge files, you need to set the purge frequency. That is the best way to handle log files. As pe...

saviop wrote:
If job 4 aborts, we have to run job 3 onwards. But this requirement is only for job 4 and not for any other job.

That statement points towards it, doesnt it

But this way even if job 5 or job 6 fails, job 3 will run, which the OP doesnt want.

What about your DS server, is that on linux? Did you try doing a google on the key words that i specified? This is more like an OS level issue than a DS issue. You need to get your linux admins involved. But i suggest you do your research first.

Split the inserts and updates
Play around with the transaction size and array size to get optimal performance.
If you are already doing that then there is something else going on in your job. Need more info on the design.

I think that would be the best way to go about doing it. Put them in one sequence job and do not make that sequence restartable.

Dicitionary like explanation Arnd. Nice
And welcome back

ray.wurlod wrote:In general, in parallel jobs, simultaneous write and read to the same object is not possible.

I think Ray has given you the answer. Avoid that and you will be ok.

I agree with Craig. Even if you try to run all 30 instances at the same time, you will exhuast the resources. Collect the files using a link collector and then process them or pass an OS level COPY command to cat the files together.

Is your DS or Teradata on linux?
It seems to be bug related to linux systems.
Heres what i found on google: "Bug#357509: free(): invalid pointer "
Try googling "*** glibc detected *** free(): invalid pointer". You will find so many results. That might help you detect the problem.

I havent personally used a Bulk loader on mainframe DB2, but i dont see why it wont. If the tables are transparent to your DB2 API, then it should work for the bulk loader as well. Why dont you go ahead and try it and let us know if it worked or not.

O yea, the sequential file properties. Totally forgot about that. Thanks Craig.

Rejection city huh, cool

Welcome Aboard Read your record as a single column and count the number of pipes, if less then 2 then send it down a reject link from the transformer. The constraint for the erraneous record will look something like DCOUNT(in.Link,"|") < 2 and for the good records it will look like...

The bulk loader stage does look intimidating with so many properties to set, but really, there are only a few that you need to worry about. A few off of my head are the database connection properties,like server name, username and password. Then you have to specify the directory path for the temp fi...

Hmm. Never done that and hence my ignorance on the topic. But its good to know that an index can be created on a hashed file.
Great!
Thanks Craig for looking out for us