DSXchange

Tom, perhaps you could explain what your "archiving" actually consists of, so that Ray can help you when he accesses this forum.

Since there is no mechanism to restore log files, taking a direct backup is not a useful exercise. You can export the log information using dsjob -log options and can save that generated output.

I'm not sure what is going wrong, I've just written a test program and can create vectors but the makesubrecord isn't working for me, either. I'm afraid I'm out of time right now to test this further (at least for today) but it seems to be some sort of an output column definition issue.

In your original post you didn't ask a specific question or give enough information to get a solution. You are trying to get the make subrecord stage work. Start off with a simple case - perhaps 2 column types and very simple test data of 2 values each and see what is coming out and what you expect....

If they can afford the overhead to do large volume file transfer on the https protocol and stay in that technology mindset then they should look into using RTI services.

25 Million rows per file is not necessarily considered huge, many sites here deal with larger volumes on an hourly basis. Looking at the files from an ETL perspective, do you even need to merge them? Can you not process them through a DataStage job sequentially? Depending upon your specifi condition...

Jobs are not stored in editable UNIX or Windows file objects, they are stored as records in hashed files (tables) in the DS repository, which uses a derivation of the UniVerse postrelational database. You can do a project export into XML format to get all your DataStage information into an editable ...

You really shouldn't be reading those values from the internal structures. The call to DSGetParamInfo() will return the same values - and will also work at the next release while I am fairly certain that your code as it stands will cease to work. The reason I asked about the command line is that DS ...

I don't think that this information is stored in the tables. I've used the DSGetParamInfo() routine to get these values and store them elsewhere.

What is your command line?

A "hung" process is one that is waiting on some resource, perhaps one that is never going to be freed. You can check up a bit from UNIX by identifying the process and then using the truss command to see what the process is doing at a system level; this might identify what the cause of the hang is. S...

Look at your Server Job Developer's Guide pages 19-140 onwards.

catchrohit4u wrote:...path is right because other hash files are opening...

Are you using parameters? how do you know the file path(s) are right - having other ones work is no proof positive that these are correct. Take one job and one file as an example.

Does that mean when you run the job the director monitor shows no rows going down that link? Or how are you determining this?

Well, DataStage certainly has found a null somewhere, perhaps in a write to an interim hashed file somewhere. You will need to specify the job & stage as well as it's type first, and then perhaps use the builtin debugger to trace the column null value which is triggering the error.