DSXchange

If POS is fixed (i.e only M/D/R) then you can use the below logic using Join stages. 1) Splitt the Input data into 3 datasets/Files, one for each POS 2)Do a full outer join to Join any two datasets.(take Dataset M and Dataset D) 3)Do the Null handling in a transformer. 4)Do a Full outer to join the ...

You can do this in two ways

1)From the Director

Job > Cleanup Resources

Select the job/Process in the Processes list box, then click Logout.

2) If you can Identify the PID related to the Job, Use the OS command to kill the process.

I guess,doing it from the Director is a better way

Your Client configuration looks fine. We had a similar issue long back. The issue was in the connectivity (between Client and Server).
Do FTP a big Flat file from your client to server and see if it is a Network connectivity/Bandwidth issue.

Right( ("000":SrcColumn) , 3 )

I guess the below logic should work.

Input----->Remove Dups(Key: ID,MO and CA)--->Agregater (Key: ID,MO) get the Count rows--->
Transformer( Constraint: if count Rows=1)---> To Lookup(Reference Stream)


Input -----> Lookup(Key: ID and MO) ---> Ouptput

If you are trying to get a Cartesian product then use a Join Stage.
or
You can use the message handler to suppress the warning from the logs, but in my opinion it is not a best practice
or
As vmcburney said, remove duplicates from your Lookup stream