I've just done a new gentoo install. When trying to ssh from this new gentoo machine to other current gentoo machines, the DISPLAY environment variable does not get set correctly. Rather than getting set to something like localhost:10.0, it gets set to the remote hostname:0.0 This only seems to happen from the new machine to one of the old but current gentoo boxes. From the new machine to a redhat 7.3 box, it gets set correctly. On the machine where the variable does not get set, the following messages appear in the system log: PAM-env[10008]: Unknown PAM_ITEM: <XAUTHORITY> su[10008]: PAM pam_putenv: delete non-existent entry; XAUTHORITY The same problem appears to happen in reverse, from the old but current gentoo box to the new one. All the gentoo boxes have pam-0.77-r1 and openssh-3.8.1_p1-r1 Reproducible: Always Steps to Reproduce: 1. 2. 3.
This depends on your sshd_config setting , if you uncomment X11DisplayOffset 10 you'll get a DISPLAY :10:0 on connect. But usually this isn't needed.
I have the same problem, with the same log messages. I've tried X11DisplayOffset and the other X11 options (even though I don't use X11 forwarding) but they don't change anything. $DISPLAY always winds up as relay.mywork.com:0.0 which causes vim to hang on startup. When I unset DISPLAY, vim starts fine - and of course this only happens over ssh. This only seemed to have started in the past week or so, possibly after a util-linux upgrade (~x86).
Uncommenting the display offset in sshd_config (and restarting sshd) doesn't make any difference. Its not just the display offset that is getting mis-set, The wrong hostname appears in the display variable. When ssh'ing _from_ home.my-isp.net _to_ my.work.ca, DISPLAY gets set to home.my-isp.net:0.0, when it should be my.work.ca:10.0, If I go and set the DISPLAY variable correctly, the windows do get forwarded.
This is a duplicate of Bug #70585
*** This bug has been marked as a duplicate of 70585 ***
