There are two "completion" in the subcommand table: bash-completion and bash_completion. They are following:
'bash_completion': MagnumClientArgumentParser(prog='magnum bash_completion', usage=None, description=None, version=None, formatter_class=<class 'magnumclient.shell.OpenStackHelpFormatter'>, conflict_handler='error', add_help=False)
'bash-completion': MagnumClientArgumentParser(prog='magnum bash-completion', usage=None, description="Prints arguments for bash-completion.\n\n Prints all of the commands and options to stdout so that the\n magnum.bash_completion script doesn't have to hard code them.\n ", version=None, formatter_class=<class 'magnumclient.shell.OpenStackHelpFormatter'>, conflict_handler='error', add_help=False)
But the command is "magnum bash-completion", so remove "bash_completion".
There are two "completion" in the subcommand table: bash-completion and bash_completion. They are following: umentParser( prog='magnum bash_completion', usage=None, description=None, version=None, formatter_ class=< class 'magnumclient. shell.OpenStack HelpFormatter' >, conflict_ handler= 'error' , add_help=False) umentParser( prog='magnum bash-completion', usage=None, description="Prints arguments for bash-completion .\n\n Prints all of the commands and options to stdout so that the\n magnum. bash_completion script doesn't have to hard code them.\n ", version=None, formatter_ class=< class 'magnumclient. shell.OpenStack HelpFormatter' >, conflict_ handler= 'error' , add_help=False)
'bash_completion': MagnumClientArg
'bash-completion': MagnumClientArg
But the command is "magnum bash-completion", so remove "bash_completion".