I’ve had similar post-update hiccups with AIO and what consistently works for me is the routine I documented here:
My AIO update routine (with recovery steps) — https://help.nextcloud.com/t/nextcloud-aio-v10-1-1-apache-will-not-start/212805/19
Key point: I use Portainer as the primary health check to confirm that all containers (nextcloud, apache, db, redis, notify_push, etc.) come up with Status = Healthy. If anything is not healthy or stuck starting (often apache or notify_push after an update), I do this:
- Stop all containers from the AIO interface.
- In Portainer, delete all stopped containers except the
nextcloud-aio-mastercontainer. - If a specific container keeps failing or shows unhealthy, also delete its image in Portainer.
- Back in AIO, let it pull a fresh image and recreate that container.
- When everything is healthy, I clean up old unused images in Portainer.
This has repeatedly fixed cases where an image pulled during update was incomplete/corrupted and a container wouldn’t reach healthy state. The symptoms (“Apache and Notify Push not starting” and NC stuck thinking it’s an initial install) match exactly the situations where step 3–4 above resolved it for me.
One more thing: sometimes the update itself simply takes longer — especially right after a new release when a huge number of servers are updating in parallel and “Source Docker Hub” is under heavy load.
On top of that, a full Nextcloud AIO setup with all modules enabled means you’re looking at roughly 15 containers running side by side. That’s the second reason I always keep Portainer open: to see how many containers are already up and healthy, and which ones are still spinning up.