Zombie processes generated in docker container

[Charles Howes]

Is this a bug report or feature request?

• Bug Report

Bug Report

What happened:

I noticed that there are about a hundred zombie ssh processes inside our ceph-mgr container.

What you expected to happen:

No zombie processes should exist. Zombie processes in a container are usually created because the parent process is not reaping children. The problem is explained at length here: https://blog.phusion.nl/2015/01/20/docker-and-the-pid-1-zombie-reaping-problem/

One solution is described in that article; another is to add tini to the Dockerfile; there are others.

How to reproduce it (minimal and precise):

Not sure; they were created a couple of weeks ago. Some of the living ssh processes in that container look like this:

ssh -C -F /tmp/cephadm-conf-vech0nrh -i /tmp/cephadm-identity-randomstring -o ServerAliveInterval=7 \
  -o ServerAliveCountMax=3 root@1.1.1.1 \
  python3 -c "import sys;exec(eval(sys.stdin.readline()))"

Environment:

• OS (e.g. from /etc/os-release): Ubuntu 20.04.4 LTS
• Kernel (e.g. uname -a): Linux example.com 5.4.0-107-generic # 121-Ubuntu SMP Thu Mar 24 16:04:27 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
• Docker version (e.g. docker version): 20.10.7
• Ceph version (e.g. ceph -v): ceph version 16.2.7 (dd0603118f56ab514f133c8d2e3adfc983942503) pacific (stable)

# docker-show-run.sh 44848cae2bca | gawk '{a=$1};END{print a}'

"quay.io/ceph/ceph@sha256:3cd25ee2e1589bf534c24493ab12e27caf634725b4449d50408fd5ad4796bbfa"

# dockerfile-from-image "quay.io/ceph/ceph@sha256:3cd25ee2e1589bf534c24493ab12e27caf634725b4449d50408fd5ad4796bbfa" | grep -e GIT_REPO -e GIT_COMMIT

LABEL GIT_REPO=https://github.com/ceph/ceph-container.git
LABEL GIT_COMMIT=6eedd2e8a16f9976675af44b3c5d8a1a247624dc