DSXchange

Hi Ray,
Can you guess the reason why it rejected certain user, while the session created by that users where still left open. By killing those session alone will help?
If the issue is with DSR_JOBS, shouldnt it give issue to all user trying to connect?

vsi wrote:update Rpt_tbl where ever you find john in rpt_tbl

I dont find John anywhere in Rpt_tbl. I guess you need to lookup based on 3rd column right?

Or call IBM you will be given another additonal routine ExecSHSilent. Which wont display the password out to the log.
Or you can frame your own encryption logic to store a encrypted password in a file and which inturns can be read and passed to the job during runtime after decryption.

Create a userdifined varialbe PWD in Datastage adminstrator and give the default value as you project home directory say C:\Ascential\DataStage\Projects\MyProject or '.' to indicate the current home directory as default.

One of the demerit of MPP system is, it is costly when compared to other archiecture due to large hardware set. Evren memory should be seperated for each CPU. Due to the same reason it gives the performace boost.

I dont have dsadm id in the new site. Sorry that, I could not help you.

If you get the whole record set leftRec_NAME should be declared as Nullable = "Yes". (rightRec_NAME as well) inorder to avoid the warning.

7.5x2 should be for PX and 7.5.2 should be for server. Even iam not sure about the official release of this version.

Before posting a question you could have tried it.
Lookup fileset is a not a very similar to fileset, other that the name which as the fileset in it.
Lookupfileset is created by datastage inorder to make a effecient lookups. And not as the fileset which is stored in the ascii characters.

Was this the only job running in the server at both time, i.e., now and two month back?
I mean to ask, is there any additional load been added to the server now. Or any other application running at the background.
Still it is running as 10 rows/sec. - Is this for the first option you tired?

MPP shared nothing architecture. Which is more costly when compared to the similar hardware configuration of SMP. Another merit, installation and maintanance on SMP would be more easiar than MPP. Even the lisencing issue.
MPP is faster than SMP due to its demerit of Shared nothing architecture.

Clients only on windows.
Server - 7.5.2 on windows.Rest on Unix/linux.

Or read each line from the file and pass as parameter to a job which inserts into the table or use comand activity to execute comand line insert from the job sequence which inturn can be scheduled.

View should point the table while retriving data, and inturns the index.

There is no option to append. But you can read the data from lookup fileset and funnel it with the additional dataset and load it back to the same lookup fileset.