ssh check failed if manager node deployed on a backup slave server

[leafgray]

Maybe it is not recommend / typical usage. My simple environment:
hostA (master)
hostB (backup master)
hostC (slave for backup) . Used as manager node.

host A/B/C can ssh login with each other use key....
but ssh_check failed，after some debug, I found it's caused by hostC can't ssh to hostC.
relate code(SSHCheck.pm line 107):
my $command = 108 "ssh $MHA::ManagerConst::SSH_OPT_CHECK -p $src->{ssh_port} $src->{ssh_user}\@$src->{ssh_ip} \"ssh $MHA::ManagerConst::SSH_OPT_ CHECK -p $dst->{ssh_port} $dst->{ssh_user}\@$dst->{ssh_ip} exit 0\""

add hostC pub key to self authorized_keys, then it works.

I don't know perl too much, there is a way to skip this situation?
or a document improvement should be ok....

Thanks.