DSXchange

Dump the score and find out where the inserted buffer(3) is to be found. This may help to diagnose what's happening. Are newTime and startTime variables or column names in your job design?

The dsenv script is to be found in your /dstage/Ascential/DataStage/DSEngine directory. Make sure that not only /dstage/Ascential/DataStage/DSEngine/lib but also /dstage/Ascential/DataStage/PXEngine/lib are in your LD_LIBRARY_PATH, and precede any database libraries in the shared library search list.

You need to specify in what column the level 01 item is. Try 8 (a common start column).

However, what you have posted is NOT a COBOL file definition. A CFD is pure text. Does it need to be converted from EBCDIC?

To aid future searchers, please post HOW this issue was resolved.

Connect via telnet, cd to the DataStage Administrator's directory (DSEngine), source the dsenv script, then invoke the DataStage shell. Execute the following queries and post the results. SELECT * FROM UV.ACCOUNT WHERE @ID = '<<ProjectName>>'; SELECT * FROM UV_SCHEMA WHERE @ID = '<<ProjectName>>'; V...

Ask your network administrator about packet size. It's usually some small multiple of 1KB.
Row size must be calculated by summing the size of each field and adding one byte for the inter-field delimiters.

Iconv() doesn't need the format but, if given, it must be right. You have omitted "A" for the month. Oconv(Iconv(ToT1.START_DATE,"D-DMY[2,A3,2]"),"D-YMD[4,2,2]") Easier to maintain would be Oconv(Iconv(ToT1.START_DATE,"D-DMY"),"D-YMD[4,2...

It is probably something in the target database. Are you sending all rows as a single transaction? Make a copy of the job and replace the database stage with a Sequential File stage, and see whether that design produces the same behaviour. If it does not, there is something in the database (or the c...

Maybe that job was still running. Two batches of 13 doesn't equal 20 - maybe your run time calculations were similarly faulty? And every five seconds suggests that the jobs all need to finish very quickly indeed.

Different issue. Please start a new thread. Search for DSJE.BADVALUE in the meantime.

Windows 2003 Server, sp1. 7.5GB memory (!), four dual-core CPUs (not sure of speed of these). Spare capacity of all kinds, though only 12GB free on disk where Projects installed. The annoying thing was, having re-booted, that NOTHING else was happening at the time.

Not easy to do in a server job. Indeed, you would need to "hack" the underlying structures, which is never to be recommended.

If you want to persist, search the forum for STAGECOM.ARR

GENERAL is a totally general-purpose algorithm that attempts to get "random" (= "flat") distribution of records over the available groups. SEQ.NUM is biased towards numeric characters, and operates right-to-left (much like an odometer). It works best when the keys form an unbroken integer sequence.

In a job sequence you can call a routine OR a command/script/BAT file. Provided it's upstream of your Job activity, its return value can be used to supply the parameter value.

It's usually memory. Try not to have too many other things open.