Hi,
I would be very careful about doing this, the process might be idle, but child processes might not be.
If for example the child process if changing a database (i.e INSERT, DELETE) and the application code has been written such that the child process recognises the parents loss and dies also. If transaction rollback is not implemented, then the database might become corrupt.
If the code has been written such that the child process ignores the loss of it's parent, then when it is finished, it will become a zombie, which can only be rid of by a reboot.
Is this the view of anyone else??
Regards
Unixqueen