“docker exec”命令的“-i”和“-t”选项的用途是什么？

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 ")