Issues using NET USE on W2K3 Servers

I submit jobs via SQL Scheduler on a W2K3 server.

SQL Job is submitted.

Job consists of program that dynamically builds a batch script which
includes issuing DOS NET USE d: \\remoteserver\sharename [my credentials]
and /persistent:no. At the end the batch script issues a NET USE d: /delete
command. Output shows all commands working fine.

Logging on to the remote server and issuing a NET USE to list the mapped
drives shows only local drives, no mapped drives.

In the past, if the job was run multiple times (but non-concurrent) it would
simply reuse the I: drive as that is the first one available. Running 2
jobs concurrently has been shown to use I: and J: drive letters and drive I:
used subsequently. So far so good.

Symptoms I'm seeing now, is I added a feature such that when the batch
script is built, it adds a NET USE to map a drive at the beginning, runs the
scripts it needs to run, all invoking CALL sub-batch.cmd, then issuing a NET
USE to delete mapping at the end. This all seems to be working. Logging on
to remote server shows no remotely mapped drives. With each subsequent run,
the next available drive letter is used, as if it thinks the server still
has the prior drives mapped.

I have looked at task manager thinking each CMD task started is hung and
retaining its' drive letter, but nothing. Task Manager looks fine. I feel
pretty confidewnt I am executing paired NET USE to map drive and NET USE to
disconnect drive. All my testing and troubleshooting is in support of that.

I can only see C:, D:, E:, F:, G:, H: as locally mapped, but last invocation
of my process showed drive Q: was selected. Again, drives I: thru P: are
NOT showing up as mapped drives opn that remote server.

Any thoughts?

regards,

doug