Like most docker users, I periodically need to connect to a running container and execute various arbitrary commands via bash.
I'm using 17.06-CE with an ubuntu 16.04 image, and as far as I understand, the only way to do this without installing ssh into the container is via docker exec -it <container_name> bash
However, as is well-documented, for each bash shell process you generate, you leave a zombie process behind when your connection is interrupted. If you connect to your container often, you end up with 1000s of idle shells -a most undesirable outcome!
How can I ensure these zombie shell processes are killed upon disconnection -as they would be over ssh?
One way is to make sure the linux init process runs in your container.
In recent versions of docker there is an --init option to docker run that should do this. This uses tini to run init which can also be used in previous versions.
Another option is something like the phusion-baseimage project that provides a base docker image with this capability and many others (might be overkill).
Related
In my CentOS server I use docker created a container,
I opened two sessions connected to the container by command:
docker attach container-name
but there is an issue, in each window I execute command the other window is display the same information.
so I cannot control the container when it is installing package.
is it possible to avoid this issue?
The docker attach command attaches to the currently running process as defined by CMD. You can attach as many times as you want, but they all connect to the same process.
If you want to access the container and have different sessions to it, use:
docker exec -it container-name bash
Or whatever shell is available. bash is common, but you may need to use sh or find out what's used, if any is there at all. Some containers are super stripped down.
The -it flag enables "interactive" mode, as otherwise it just runs that command and shows you the output.
Running docker run with a -d option is described as running the container in the background. This is what most tutorials do when they don't want to interact with the container. On another tutorial I saw the use of bash style & to send the process to the background instead of adding the -d option.
Running docker run -d hello_world only outputs the container ID. On the other hand docker run hello_world & still gives me the same output as if i had run docker run hello_world.
If I do the both experiments with docker run nginx I get the same behavior on both (at least as far as I can see), and both show up if I run docker ps.
Is the process the same in both cases(apart from the printing of the ID and output not being redirected with &)? If not, what is going on behind the scenes in each?
docker is designed as C-S architecture: docker client, docker daemon(In fact still could fined to container-d, shim, runc etc).
When you execute docker run, it will just use docker client to send things to docker daemon, and let daemon call runc etc to start a container.
So:
docker run -d: It will let runc run container in background, you can use docker logs $container_name to see all logs later, the background happend on server side.
docker run &: It will make the linux command run at background, which means the docker run will be in background, this background run on client side. So you still can see stdout etc in terminal. More, if you leave the terminal(even you bash was set as nohup), you will not see it in terminal, you still need docker logs to see them.
I pulled centos6 image and made a container from it. I got its bash by:
$ docker run -i -t centos:centos6 /bin/bash
On the centos6 container, I could use "service" command without any problem. But when I pulled&used centos7 image:
$ docker run -i -t centos:centos7 /bin/bash
Both of "service" and "systemctl" didn't work. The error message is:
Failed to get D-Bus connection: Operation not permitted
My question is:
1. How are people developing without "service" and "systemctl" commands?
2. If I want to use, for example, httpd.service on the centos7 container, what should I do? Or maybe running services on a container is not recommended?
There is no process supervisor running inside either container. The service command in your CentOS 6 container works by virtue of the fact that it just runs a script from /etc/init.d, which by design ultimately launch a command in the background and return control to you.
CentOS 7 uses systemd, and systemd is not running inside your container, so there is nothing for systemctl to talk to.
In either situation, using the service or systemctl command is generally the wrong thing to do: you want to run a single application, and you want to run it in the foreground, so that your container continues to run (from Docker's perspective, a command that goes into the background has exited, and if that was pid 1 in the container, the container will exit).
How are people developing without "service" and "systemctl" commands?
They are starting their programs directly, by consulting the necessary documentation to figure out the appropriate command line.
If I want to use, for example, httpd.service on the centos7 container, what should I do? Or maybe running services on a container is recommended?
You would start the httpd binary using something like:
CMD ["httpd", "-DFOREGROUND"]
If you like to stick with service/sytemctl commands to start/stop services then you can do that in a centos7 container by using the docker-systemctl-replacement script.
I had some deployment scripts that were using th service start/stop commands on a real machine - and they work fine with a container. Without any further modification. When putting the systemctl.py script into the CMD then it will simply start all enabled services somewhat like the init-process on a real machine.
systemd is included but not enabled by default in CentOS 7 docker image. It is mentioned on the repository page along with steps to enable it.
https://hub.docker.com/_/centos/
If I use docker exec to fire up a shell,
docker exec -ti <CONTAINER> /bin/bash
I could use Ctrl+p Ctrl+q to detach this shell process. Then this shell is still running inside the container, but how can I reattach to that one particular shell (the one started by docker exec, not docker run)?
Sadly, this is not possible yet; see this issue on GitHub. I've also wanted this functionality, but at the moment it seems like there's no direct way to do this.
A workaround has been proposed, to take care of the case where you're accessing a box via ssh and running docker exec on the remote box (or, for the case where your terminal emulator is unstable and may crash on you): Always run your docker exec commands inside screen or tmux. If you do this, whenever you get detached from the screen/tmux session, you can re-attach to it later and still have your docker exec commands accessible. (this is a bit different than what was suggested by #vodolaz095, since it involves running screen or tmux outside the container, making it suitable for use with containers that don't run screen/tmux as their main process)
docker exec is specifically for running new things in an already started container, be it a shell or some other process.
docker attach is for attaching to a running process, so you can use only one instance of shell.
Run you container(process)
docker run -tid --name <CONTAINER> <IMAGE>:<TAG> bin/bash
Then
docker attach <CONTAINER>
To detach Ctrl+p + Ctrl+q
On this way you can attach and detach multiple times with only one instance of shell
I have been running docker processes (apps) via
docker run …
But under runit supervision (runit is like daemontools) - so runit ensures that the process stays up, passes signals etc.
Is this reasonable? Docker seems to want to run its own demonization - but it isn't as thorough as runit. Furthermore, when runit restarts the app - a new container is created each time (fine) but it leaves a trace of the old one around - this seems to imply I am doing it in the wrong way.
Should docker not be run this way?
Should I instead set up a container from the image, just once, and then have runit run/supervise that container for all time?
Docker does do some management of daemonized containers: if the system shuts down, then when the Docker daemon starts it will also restart any containers that were running at the time the system shut down. But if the container exits on its own or the kernel (or a user) kills the container while it is running, the Docker daemon won't restart it. In cases where you do want a restart, a process manager makes sense.
I don't know runit so I can't give specific configuration guidance. But you should probably make the process manager communicate with the docker daemon and check to see if a given container id is running (docker ps | grep container_id or equivalent, or use the Docker Remote API directly). If the container has stopped, use Docker to restart it (docker run container_id) instead of running a new container. Or, if you do want a new container each time, then begin with docker run -rm to automatically clean it up when it exits or stops.
If you don't want your process manager to poll docker, you could instead run something that watches docker events.
You can get the container_id when you start the container as the return value of starting a daemon, or you can ask Docker to write this out to a file (docker run -cidfile myfilename, like a PID file)
I hope that helps or helps another runit guru offer more detailed advice.
Yes, I think running docker under runit makes sense. Typically when you start a process there is a way to tell it not to daemonize if it does by default since the normal way to hand-off from the runit run script to a process is via exec on the last line of your run script. For docker this means making sure not to set the -d flag.
For example, with docker you probably want your run script to look something like this:
#!/bin/bash -e
exec 2>&1
exec chpst -u dockeruser docker run -a stdin -a stdout -i ...
Using exec and chpst should resolve most issues with processes not terminating correctly when you bring down a runit service.