[Bug 3648] ssh client blocks itself by leaving unix domain socket alive after user

bugzilla-daemon at mindrot.org bugzilla-daemon at mindrot.org
Sun Mar 10 10:12:03 AEDT 2024


https://bugzilla.mindrot.org/show_bug.cgi?id=3648

--- Comment #3 from Hadmut Danisch <hadmut at danisch.de> ---
Apparently, yes, partly, thanks. It seems to work for  LocalForward,
but not for RemoteForward. The option exists in the sshd_config man
page, so I guess this must be set on the server side as well in order
to make it work for RemoteForward, but that doesn't help in that
particular use case, since users do not have admin rights on the remote
machine.  



Didn't see that option, and I do know about other long-time-experienced
admins with the same problem, who didn't see this either. 

Maybe a hint in the LocalForward and RemoteForward sections of man page
for ssh_config would be useful. 



But what sense does it make to leave the socket in the filesystem, if
ssh will never ever again be able to reuse it, it just keeps ssh from
working again? What's that good for?


I'd see two modes of operation: 

- either leave the socket intact, but be able to reuse it with
subsequent ssh sessions, which might make sense for other client
programs using it, 

- or remove it and create a new one for every ssh connection. 



but the current default mode of ssh does not seem to be working and
thus not seem to make sense. 


(BTW: Did you see my e-mail to the mailing list about XDG_RUNTIME_DIR
environment variable on the server side?)


regards

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.


More information about the openssh-bugs mailing list