Existing ExApps can end up stuck in a broken state after Docker Proxy → HaRP transition

[vawaver]

Describe the bug

Hi,

I would like to report an issue affecting already deployed ExApps after moving from Docker Proxy for ExApps to HaRP.

I hit this after updating to Nextcloud AIO 12.9.1 Beta on the Beta channel, where I switched from Docker Proxy for ExApps to HaRP.

Problem

I previously had these ExApps deployed:

• Nextcloud Context Agent
• Assistant Talk Bot

After the update, the old ExApp containers were still present and running in Docker, but from the Nextcloud/AppAPI side they were no longer handled correctly.

From the user perspective, this resulted in a broken in-between state:

• Deploy and Enable was inactive
• Docker Proxy showed as unavailable
• the apps could not be properly disabled, re-deployed, or recovered from the GUI
• the old ExApp containers were still running in Docker/Portainer
• app_api:app:list showed no registered ExApps

So effectively, the old ExApps were still running as containers, but AppAPI no longer had a usable state for managing them.

Observed behavior

What I observed was:

• old ExApp containers still existed and were running
• AppAPI no longer listed them as registered ExApps
• the GUI controls were no longer usable
• manual cleanup was needed before the apps could be deployed again through HaRP

In practice, this meant I had to:

• manually remove the old leftover ExApp containers
• manually re-register the deploy daemon
• then manually redeploy the ExApps from the GUI

Expected behavior

I would expect already deployed ExApps to be handled more gracefully during this transition.

For example, one of these behaviors would be much clearer from the user perspective:

• existing ExApps are migrated automatically if possible
• existing ExApps are clearly marked as needing redeployment
• the GUI offers a recovery path instead of leaving inactive controls
• AppAPI provides a cleaner state instead of “containers still running, but no registered ExApps”

Actual behavior

The system ended up in a half-migrated state where:

• old ExApp containers were still present
• AppAPI no longer had a usable registration state for them
• the GUI could not manage them anymore
• the user had to figure out the cleanup manually

Additional note

During my recovery, I also ran into a 401 Unauthorized error while checking the new HaRP connection. In my case, this was likely my own mistake because I had overwritten the shared key shown in the GUI and later corrected it with the actual HP_SHARED_KEY from the running nextcloud-aio-harp container.

So I do not want to overstate that part as an AppAPI bug.

The main issue I want to report here is the handling of already deployed ExApps during the transition itself.

Environment

• Nextcloud AIO 12.9.1 Beta

• Beta channel

• migration from Docker Proxy for ExApps to HaRP

• affected ExApps:

  • Nextcloud Context Agent
  • Assistant Talk Bot

Steps/Code to Reproduce

all is mentioned above

Expected Results

all is mentioned above

Actual Results

all is mentioned above

Setup configuration

all is mentioned above