New answers tagged sshd
Is the server configured to accept RSA keys? Look for RASAuthentication in the sshd_config file.
The $HOME/.ssh directory mode must be 700 and authorized_keys should be readable by the owner only, i.e. mode 600: chmod u=rwx,g=,o= /var/git/.ssh chmod u=rw,g=,o= /var/git/.ssh/authorized_keys As to the private key, it must be read-write-able by the owner only: chmod u=rw,g=,o= /var/git/.ssh/id_?sa
For reasons of paranoia, the .ssh directory and authorized_keys must not be group-writable. I guess the thinking is, the user must be the only one with explicit control over his/her authorization. I believe a work-around for this lies with ACL. The other work around is StrictModes=no setting in sshd's configuration file. But it would be too dangerous to do ...
The problem is the fact that files permissions are too open. Try setting the mode of authorized_keys to 600 and the .ssh directory to 700.
Most importantly you should have permissions to do so. Most of the problem like cannot start process or cannot find some file are due to permissions. Use sudo before any command. Now for ssh you can simply do sudo stop ssh sudo start ssh This leverages upstart
On Solaris 11 you can use: vi /etc/ssh/sshd_config PermitEmptyPasswords yes Reload sshd Add a new user useradd -m -d /home/testuser -s /bin/bash testuser passwd -d testuser SSH using new user admin@testhost:~$ ssh testuser@localhost Last login: Tue May 19 15:02:42 2015 from localhost Oracle Corporation SunOS 5.11 11.2 December 2014 ...
This is an SELinux problem, caused by your home directory being in a weird location. The thing you did to disable it, didn't. SELinux isn't a service, and systemctl status selinux is just telling you that ("not-found (Reason: No such file or directory)"). You could run setenforce permissive or otherwise disable SELinux, but this kind of like taking the ...
It's SELinux. Even though I stopped and disabled it using systemctl, I still needed to use /usr/bin/setenforce 0. Passwordless login now works as expected.
My issue was with the /etc/resolv.conf - wrong domains to search upon, as well as non-existent DNS servers to provide lookups. Surprised I was able to login at all.
Did you restart the ssh service? Are two consecutive SSH logins (within a minute or two) both slow? Perform ssh -vvv user@host from the client to see where it waits. If that does not help, you might stop the SSH service (if you have other ways to access the host) and start SSH in the foreground with debugging on: sshd -Dd. Then try to make a SSH connection ...
General troubleshooting advice for OpenSSH First of all I refer you to this short troubleshooting guide for sshd which I am using as a recipe time and time again. The plot thickens Only difference in this case, I used lxc-console to attach to the guest, logged in and stopped the running sshd and then started my instance on the default port 22. And then I ...
Rather than using Match, if you wish to allow logging in from a single host, the following works for me (in sshd_config): AllowUsers *@192.168.0.4 It only allows users logging in from 192.168.0.4, using any login on the target. You can replace * with a specific login if you wish, and specify multiple patterns separated by spaces; so for example: ...
Top 50 recent answers are included