linux - What is the purpose of the "-i" and "-t" options for the "docker exec" command?

Question

To be honest, I have always been confused about docker exec -it …, docker exec -i … and docker exec -t …, so I decide to do a test:

1. docker exec -it …:

   # docker exec -it 115c89122e72 bash
   root@115c89122e72:/# ls
   bin  boot  dev  etc  home  lib  lib64  media  mnt  opt  proc  root  run  sbin  srv  sys  tmp  usr  var
   

   It works normally.

2. docker exec -i …:

   # docker exec -i 115c89122e72 bash
   ^C
   

   The command hangs and I have to use Ctl + c to interrupt it.

3. docker exec -t …:

   # docker exec -t 115c89122e72 bash
   root@115c89122e72:/# ls
   ^C
   

   It enters the container successfully but hangs on executing the first command.

So it seems there is no point in having the docker exec -i … and docker exec -t … commands. Could anyone elaborate on why there exist -i and -t options for the docker exec command?

Answer 1

-i, --interactive keeps STDIN open even if not attached, which you need if you want to type any command at all.

-t, --tty Allocates a pseudo-TTY, a pseudo terminal which connects a user's "terminal" with stdin and stdout. (See container/container.go)

If you do an echo, only -t is needed.
But for an interactive session where you enter inputs, you need -i.

Since -i keeps stdin open, it is also used in order to pipe input to a detached docker container. That would work even with -d (detach).
See "When would I use --interactive without --tty in a Docker container?":

$ echo hello | docker run -i busybox cat
  hello

---

-i keeps STDIN open even if not attached, what is the status of STDOUT in this case?

It is, for docker exec, the one set by docker run.

But, regarding docker exec, there is a current issue (issue 8755: Docker tty is not a tty with docker exec

unfortunately your discovery only amounts to a difference between the behaviour of tty in centos6 vs ubuntu:14.04. There is still not a functional tty inside the exec - just do ls -la /proc/self/fd/0 and see that it's a broken link pointing to a pts which doesn't exist.

the actual bug we're dealing with is that certain standard libraries assume that the symlinks in /proc/self/fds/ must be valid symlinks

The problem is that the tty is created outside on the host and there is no reference to it in the container like how /dev/console is setup in the primary container.
One options to fix this would be allocate and bind mount the devpts from the host in to the containers.

Note (Q4 2017): this should been fixed by now (docker 17.06-ce).
See PR 33007.

That PR now allows (since 17.06):

zacharys-pro:dev razic$ docker run --rm -t -d ubuntu bash
83c292c8e2d13d1b1a8b34680f3fb95c2b2b3fef71d4ce2b6e12c954ae50965a

zacharys-pro:dev razic$ docker ps
CONTAINER ID        IMAGE               COMMAND             CREATED             STATUS              PORTS               NAMES
83c292c8e2d1        ubuntu              "bash"              2 seconds ago       Up 1 second                             xenodochial_bardeen

zacharys-pro:dev razic$ docker exec -ti xenodochial_bardeen tty
/dev/pts/1

(before 17.06, tty was returning "not a tty")