Pty Allocation Problem

So far, this problem has mostly occurred on server-1 on both clusters, probably because it gets the most login traffic suring maintenance. It seems to be caused by some combination of entering chroot, or being autologged out, or maybe emacs shell mode seems to use up PTY devices; these are assocatied with interactive shell sessions. When this occurs it becomes impossible to start an xterm, either via remote SSH or even from the GUI desktop (i.e., on the console). Rebooting fixes this, but below is a fix that restores it without resorting to a reboot.

  • To gain access to a terminal running the shell do ssh server-1 /usr/bin/bash -i from server-2 or probably via anywhere else that allows sshing into server-1.
  • From the command line do umount /dev/pts; mount -t devpts devpts /dev/pts
  • This issues lots of warnings that sound like the command did not work. However, it has so far.
  • At some point it would be nice to prevent this from happening or to at least be able to detect it short of being locked out.

-- JimJacobs - 2020-09-21
Topic revision: r1 - 2020-09-21, JimJacobs
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback