dockerfile ssh zombie process for ssh autostart

Hi I have some huge problems with autostart sshd with container.

My dockerfile: Dockerfile

My entrypoint: entrypoint

When I enter in to container bash, and type:

service ssh status

[FAIL] sshd is not running ... failed!

also I get zombie ssh process :-|

ps -ef | grep ssh

node        15     1  0 14:49 ?        00:00:00 [sshd] defunct

node       183   142  0 14:59 pts/0    00:00:00 grep ssh

Did I make some mistakes in the dockerfile ??

This is because you use USER node at the end of Dockerfile to start sshd, which I guess you want to use node user to start npm.

But, the suggested way is to use root to start sshd & use node to start npm, you can see a famous project redis which use same solution here

Then, you needed to next fix:

1. Delete USER node at the end of Dockerfile before CMD.

2. Delete RUN chmod 0444 /etc/ssh/* in your dockerfile

   Otherwise, it will reported next which make sshd not work:

   Permissions 0444 for '/etc/ssh/ssh_host_ecdsa_key' are too open.

3. Delete RUN echo 'PermitRootLogin=without-password' >> /etc/ssh/sshd_config, use next to replace:

   RUN echo 'PermitRootLogin=yes' >> /etc/ssh/sshd_config
   

4. Add RUN apt-get install -y gosu in Dockerfile to install gosu which will later be used in entrypoint.sh

5. In entrypoint.sh, change exec "$@" to next:

   exec gosu node "$@"
   

   This will assure npm start still run with user node.

Then, you can see when start the container, the sshd works, you can use service ssh stop && service ssh start to restart the service if you needed, but as the container run sshd well now, I guess you no need to use this again.