[THREESCALE-12007] run system-app with default cmd

[akostadinov]

Assisted-by: Claude Code

not tested at all, just an example

[briangallagher]

✅ Snyk checks have passed. No issues have been found so far.

Status	Scanner	Critical	High	Medium	Low	Total (0)
✅	Licenses	0	0	0	0	0 issues
✅	Open Source Security	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.

[tkan145]

@akostadinov Unfortunately, this doesn't work.

• Deploy 3scale with 2.16 image

image: 'registry.redhat.io/3scale-amp2/system-rhel9@sha256:f9926c0b1131cf99a491402251a985bc334415b4bd5cad076adaff0826c45e9f'

• Attach debug pod

kubectl debug -it system-app-769b8bbc76-6v9kv --image=registry.redhat.io/rhel8/support-tools --target system-master

• Check running command

bash-4.4$ ps aux                                                                                                                                                                       
USER         PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND                                                                                                             
1000680+       1  0.0  0.0   1088   512 ?        Ss   06:28   0:00 /usr/libexec/catatonit/catatonit -- /opt/system/entrypoint.sh unicorn -c config/unicorn.rb -E ${RAILS_ENV} config.ru
1000680+       2 10.7  1.3 427128 278620 ?       Sl   06:28   0:13 unicorn master -c config/unicorn.rb -E ${RAILS_ENV} config.ru                                                       
1000680+      43  0.0  1.2 433284 258724 ?       Sl   06:28   0:00 unicorn worker[0] -c config/unicorn.rb -E ${RAILS_ENV} config.ru                                                    
1000680+      44  0.0  1.2 433284 258552 ?       Sl   06:28   0:00 unicorn worker[1] -c config/unicorn.rb -E ${RAILS_ENV} config.ru                                                    
1000680+      57  0.2  0.0  12140  3200 pts/0    Ss   06:30   0:00 /usr/bin/bash                                                                                                       
1000680+      58  0.0  0.0  44788  3328 pts/0    R+   06:30   0:00 ps aux                                                                                                              

• Check RAILS_ENV

bash-4.4$ echo $RAILS_ENV
                         

I like to explicitly set the arguments inside the operator, it's easier to track what was used rather than having to dig into the container image or attach a debug pod

[akostadinov]

RAILS_ENV should be set within the image itself. I don't think the processes would work if it was not set and I don't think the PR changes how RAILS_ENV is set in any way.

I think it is normal that you don't see that variable using a debug pod of a totally unrelated image.

I think you need to look also for TENANT_MODE and PORT variables. The source of truth should be /proc/<pid>/environ and you can checked it also by podman exec or in OCP open pod terminal, no need for a separate debug pod.

I prefer that we update the image and then it run in the mode that we want instead of having discrepancies between image definition and runtime environment. Less moving parts. That's my vote.