I have several local machines that, while working, I frequently use ssh
to do something on a different machine. While on that second machine,
I may even use ssh to do something on a third machine. All of this
works exactly as it should. FWIW, all of my local machines run Fedora.
Fedora currently includes openssh-9.6p1-1.fc40.4.x86_64 package.
If I use JuiceSSH on my Android phone to contact one of my local
machines, that also works well. But if, while connected, I try
to start an ssh session to a different local machine, it simply
times out. The last message seen with "ssh -vvv" is, "debug3:
ssh_get_authentication_socket_path:". Up to that point, the output
of "ssh -vvv" is identical when running successfully between two
local machines without using JuiceSSH.
If instead of trying to start an ssh session immediately upon
connecting with JuiceSSH, I first use sudo to become root on the
local machine and then try to ssh to another local machine, that works
perfectly (provided, of course, that the second machine permits root
ssh login). The second ssh session also works if I simply run "sudo
su {me} -c ssh" so becoming root isn't the key.
Summary:
JuiceSSH -> user@machine1 -> user@machine2 :: FAILS
JuiceSSH -> user@machine1 -> sudo on machine1 -> user@machine2 :: WORKS
I've examined the output of the set command for all of these situations
and can't find any variables that are different that seem relevant. (Of
course, $PPID and similar variables are different.) Can anyone suggest
what could be the issue? Or what other tests might be helpful?
--
Dave Close, Compata, Irvine CA +1 714 434 7359
dave@xxxxxxxxxxx dhclose@xxxxxxxxxxxxxxxxxx
"Change is good. You go first." -- Dilbert
_______________________________________________
openssh-unix-dev mailing list
openssh-unix-dev@xxxxxxxxxxx
https://lists.mindrot.org/mailman/listinfo/openssh-unix-dev
