Problem: Zombie processes on host after in update 2.9.109

[guyke01]

Issue

Since update 2.9.109 i see a lot of zombie processes on the Ubuntu host form SSH after running a playbook. It seems that the proces is not exiting after completing a runbook.

In the previous version 2.9.75 this behavior was not occuring. The logs provided are the zombie processes after running a playbook. After restarting the semaphore container the zombie processes are killed.

Impact

Ansible (task execution)

Installation method

Docker

Database

MySQL

Browser

Microsoft Edge

Semaphore Version

2.9.109

Ansible Version

N/A. Docker installation.

Logs & errors

id ansible
uid=1001(ansible) gid=1001(ansible) groups=1001(ansible)

docker inspect $(docker ps -q) --format '{{.Config.User}} {{.Name}}'
1001 /semaphore_server
 /semaphore_db

ansible     46284  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46284  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46285  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46286  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46287  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46289  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46290  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46291  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46292  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46294  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46295  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46296  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46297  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46299  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46300  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46301  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46302  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46304  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46305  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>
ansible     46306  0.0  0.0      0     0 ?        Zs   18:00   0:00 [ssh] <defunct>
ansible     46307  0.0  0.0      0     0 ?        Z    18:00   0:00 [ssh] <defunct>

pstree -p -s 46307
systemd(1)───containerd-shim(34474)───semaphore(34520)───ssh(46307)

docker compose restart
[+] Restarting 2/2
 ✔ Container semaphore_server  Started                                                                                                                              0.5s 
 ✔ Container semaphore_db      Started                                                                                                                              2.0s

ps aux | grep 'Z'
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root       51865  0.0  0.1   6480  2280 pts/0    S+   21:47   0:00 grep --color=auto Z

Manual installation - system information

No response

Configuration

No response

Additional information

No response

[tboerger]

Potentially because tini have been removed. Tini prevents zombie processes.

[fiftin]

Hi @guyke01 , @tboerger I reproduces the issue. Yes, it is true, tini kills zombies. Will be returned in next release (ASAP).