Error " unable to save state file: a session with PID 25992 is already running on platform "cluster" " with odo v3.11.0 using Openshift local

[jochenr]

Hello,

using odo (from vscode's openshift tooling) I sporadically/very often get this error:

Starting odo dev --debug
__
/ __ Developing using the "quarkus-camel-xxx" Devfile
_/ \ Namespace: xxx-dev
/ _/ odo version: v3.11.0
__/

• Running on the cluster in Dev mode
  X unable to save state file: a session with PID 25992 is already running on platform "cluster"

Press any key to close this terminal

Environment is
Openshift local
CRC version: 2.20.0+f3a947
OpenShift version: 4.13.0

running on Windows 10

I am not able to finde the cause of the problem.

I'd appreciate some help.

Thank you,
Jochen

[rm3l]

Hi @jochenr,

From the error message, it looks like there is already an "odo dev" session running on your system for this component. odo only supports one running Dev Session per platform type (cluster or podman).

Can you please check if there is any running process with the PID reported by odo (25992 from the output you shared)?
Maybe we have a bug in the way we detect already running processes on Windows.

A workaround can also be to restart the operating system.

[jochenr]

Hi @rm3l ,

thank you for the hint.
This morning, when I experienced the issue again, I took a look on the windows task manager and didn't find the PID.
Then I used "get-process" in powershell and found it.

The strange thing is that the PID belonged to a completely other program I was running (https://conemu.github.io/ a terminal window running my AMQ Artemis instance).

After killing this process, ODO worked fine.

But why does ODO try to use a PID that is already used by another program?

[rm3l]

Thanks for the detailed information!

But why does ODO try to use a PID that is already used by another program?

When starting, odo dev stores its own PID into an internal .odo/devstate* file (in the project working dir). Then if you try to run odo dev again against a same platform, it uses this file and checks whether there is already a running process with that PID. This file should be deleted when odo dev is stopped.
After discussing this with the rest of the team, the problem we suspect here is that the odo dev session might not be properly stopped by the VSCode Plugin on Windows, and the previous PID recorded was assigned to a different process by the operating system (but we'll need to confirm this and report the issue). And we plan to create an issue to make odo delete component able to properly delete all these internal state files.
Meanwhile, I think a simpler workaround for you can be to manually delete all the .odo/devstate* files from the project working directory (where your devfile.yaml is located) before starting odo.

[jochenr]

Thank you very much!👍

The workaround helped😃