I don't know if it is possible to run user (i.e. client) requested backups of NetApp filers via NDMP. To try I guess you could create a timed release (scheduled) policy, and get that working first. And then try creating a user mode schedule with the required session window, and try requesting a run of that schedule from the command line.
Here's a normal "user mode" (i.e. client requested) user schedule bpbackup example:
$ set z_nbu=C:\test.nbu
$ set z_log=C:\test.log
$ bpbackup -p ADMIN-CONSOLES -h %computername% -s DAILY_USER -t 13 -L "%z_log%" -w -f "%z_nbu%"
Example list of files “C:\test.nbu”
7 C:\TEST
bpbackup
-p policy-name Defines storage unit, and thus media server to use.
-h client-name If left blank, it will run the policy for all clients defined within the policy.
-s schedule-name This controls retention level.
-t policy-type (13 = Windows, 0 = standard Unix OpenVMS)
-L log-file Progress log file
-w hh:mm:ss Time to wait for completion.
-f "file.ext" Specifies a file which contains a list of files to be secured, in the form <len><space><path> where <len> is the length of the path, not the length of the line/record.
Notes:
The log file is appended to. If a fresh log file is required for each run, then delete it before requesting a backup. The log file and "file list" file (i.e. "test.nbu") must exist on a local drive visible to the "system" account. Do not use mapped drives, as this causes error 12 (file read failure) and also causes backup jobs to hang. The log file location has to be a path that is visible to the account that runs the actual NetBackup job.
Other Switch examples:
-w without a parameter, is the same as -w 0
-w 0 wait indefinitely
-w hh:mm:ss maximum wait is 23:59:59
(if the time expires, then the backup continues being saved to the media server, i.e. it is NOT terminated/aborted, but the local "bpbackup" command does exit with a timeout status code.)
Also note that "user" requested backups are always full backups. I'm not sure whether this applies to NDMP based backups of NetApp filers.
If you find that the above doesn't work, another alternative is to get a client/media/master to fire an "immediate" backup, simply by requesting a normal backup, i.e. use a normal full/cinc/diff schedule that has no actual "session window" defined, and call using e.g.:
$ bpbackup -i -S $master -h $client -p $policy -s $schedule -w 12:00:00 -L C:\log.log
...but the user/computer executing this command has to be an admin console or a client that has "server" priviledge over the master, i.e. the requesting computer name has to exist in the list of "servers" of the master (i.e. for Unix, the requester has to have a SERVER=<name> entry in the "bp.conf" on the master server). Note that this method can be used to allow clients to request "incremental" and "differential" backups - but at the expense of weakening security by allowing the client machine admistrative rights over NetBackup.