It's essential for the method to work.  ssh-agent opens a named pipe
(look in /tmp/ssh-<username>/) which allows all sub-processes to share
passwords for private keys.  Each sub-processor inherits the
environment space which includes an environment variable with the
pipe's name.  ssh-agent writes new password data to this pipe.

-d

Raffaele Cicchese writes:
 > Hi all,
 > does anyone know why ssh-add2 needs to have ssh-agent2 as ancestor
 > process? Is it a security issue?
 > 
 > Thanks
 > Raffaele.

Reply via email to