nvm.sh spawns zombie processes when running on zsh in Docker

[soareschen]

When docker exec into a Docker container with zsh as shell, nvm.shwould generate a dozens of zombie processes. This problem only arise when nvm is running on zsh. There is no zombie process when running nvm with bash.

I understand that many Docker containers do not run with a proper init process, and it is inevitable that zombie process will eventually lurk out. But I also want to avoid as much zombie process as possible and try to get nvm work well with most Docker containers.

Sample command line output:

docker exec -it $IMAGE_ID /bin/zsh
source ~/.nvm/nvm.sh
ps aux
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root         1  0.1  0.4  19388  4248 ?        Ss+  10:49   0:00 /bin/bash
root       145  0.5  0.4  34792  4720 ?        S+   10:49   0:00 /bin/zsh
root       170  0.0  0.0      0     0 ?        Z+   10:49   0:00 [zsh] <defunct>
root       175  0.0  0.0      0     0 ?        Z+   10:49   0:00 [zsh] <defunct>
root       177  0.0  0.0      0     0 ?        Z+   10:49   0:00 [zsh] <defunct>
root       182  0.0  0.0      0     0 ?        Z+   10:49   0:00 [zsh] <defunct>
root       183  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       184  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       186  0.0  0.0      0     0 ?        Z+   10:49   0:00 [zsh] <defunct>
root       191  0.0  0.0      0     0 ?        Z+   10:49   0:00 [zsh] <defunct>
root       220  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       222  0.0  0.0      0     0 ?        Z+   10:49   0:00 [grep] <defunct>
root       223  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       224  0.0  0.0      0     0 ?        Z+   10:49   0:00 [grep] <defunct>
root       225  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       226  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       227  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       228  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       229  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sort] <defunct>
root       230  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sort] <defunct>
root       231  0.0  0.0      0     0 ?        Z+   10:49   0:00 [sed] <defunct>
root       237  0.0  0.0      0     0 ?        Z+   10:49   0:00 [zsh] <defunct>
root       249  0.0  0.2  15576  2156 ?        R+   10:49   0:00 ps aux

[ljharb]

nvm does launch a large number of subshells, but I'm not sure why they'd be zombies, since they're just executed, and their output captured.

I'll certainly leave this open, but I have no idea how to fix this - does this ever happen with zsh outside of Docker?

[soareschen]

I have done more testing and confirmed that this is a bug on both zsh and Docker. So it is not really a problem of nvm. More details at https://gist.github.com/soareschen/240e49116c7f2632d179

[ljharb]

Phew, thanks for clarifying. I'll close this issue.