how to rerun clone job manually in networker 9.0.0.5

Simply run nsrclone with the appropriate parameters.

Ouch, being stuck at nw9.0.0.5? I don't envy you.

With manually you mean indeed via CLI? With the intention to run it as configured with the workflow and the query group that determines what is supposed to be cloned via the workflow, or rather as bingo.1 states, by using nsrclone, completely separate from any workflow definitions (I can recall bingo.1 prefers using nsrclone anyways above using the workflows as nsrclone is more powerful if you want to do some advanced cloning, like filtering what to clone and what not, that the NMC gui does not support or at least not as flexible)?

Are we even talking about Linux or Windows here?

On linux:
# nohup sudo nsrworkflow -p " " -w " " &

If however memory serves me well, very early nw9 versions also allowed a workflow to be started with nsrpolicy? But it is too long ago, and we never even ran nw9.0, but started out with nw9.1 or so? Can't recall, way too long ago...

I'd consider implementing a supported nw version also? nw9.0.0.5 is way too edgy to my liking, even at the time, way too many issues in those initial releases, almost completely useless to try to upgrade large(r) nw8 environments with it, as it might kill itself as it might run into issues of needing to add too many workflows into the default policy, going way beyond the recommended 50 groups (some of our nw servers would require even 2000+ groups, so only with nw18.2 that was tackled by randomly adding 50 workflows during savegroup conversion to a new policy Default, Default_1, etc)). There were some KB articles stating it would kill the NMC when selecting any policy containing too many workflows.

Just one of the examples of the abomination that early nw9.0 was.

As so often users do not state full information about their question/problem.

For the clone of a single save set i do not see nsrclone creating a problem. It is not bad to learn about NW and how it deals with/manages the save set and their respective internal index information.

However, if you have to deal with multiple save sets, using nsrworkflow might simplify your job. But in the end, a nsrclone job will be started anyway - so what is wrong by using nsrclone right away?