App tokens are unavailable because the "auth-app" service is not running

[SangeloDev]

Describe the bug

The app-token page seems to be inaccessible, showing this error message:
"App tokens are not available because the »auth-app« service is not running. Please contact an administrator."
And previously created tokens are invalid and can't be used to authenticate.

I was able to use the page and set up multiple keys normally, but at some point, this bug occured and it stopped working.
I haven't changed my configuration since initially installing OpenCloud, and I've restarted the stack multiple times already.

There are little to no logs that I found relating this, apart from this:

> docker compose down && docker compose up -d && docker compose logs -f | grep auth
...
opencloud-1  | {"level":"info","service":"auth-app","host.name":"d4bcfeeb6413","pkg":"rgrpc","time":"2026-03-26T11:24:11Z","line":"github.com/opencloud-eu/reva/v2@v2.42.5/pkg/rgrpc/rgrpc.go:192","message":"grpc server listening at tcp:127.0.0.1:9246"}
...
opencloud-1  | {"level":"error","service":"auth-app","actor":"a22aeafb-d519-498c-9da7-65153fa6a86f","status":"CODE_INTERNAL","time":"2026-03-26T11:24:52Z","line":"github.com/opencloud-eu/opencloud/services/auth-app/pkg/service/service.go:196","message":"error listing app passwords"}
opencloud-1  | {"level":"info","service":"proxy","proto":"HTTP/1.1","request-id":"d4311791-9489-4b11-8568-221fb3e7707d","traceid":"0b95a6c766d1966cd9b3f40a7989aa47","remote-addr":"<IP_ADDRESS>","method":"GET","status":500,"path":"/auth-app/tokens","duration":217.58232,"bytes":0,"time":"2026-03-26T11:24:52Z","line":"github.com/opencloud-eu/opencloud/services/proxy/pkg/middleware/accesslog.go:34","message":"access-log"}
...

Steps to reproduce

1. Unknown, happened randomly

Expected behavior

The page should be accessible and valid tokens should be usable.

Actual behavior

The page is inaccessible and valid, previously generated tokens, are not usable. They throw a 401 error.

Setup

Details

Relevant configuration options:

COMPOSE_FILE=docker-compose.yml:compose-override.yml:radicale/radicale.yml:antivirus/clamav.yml:search/tika.yml:external-proxy/opencloud.yml
OC_DOCKER_IMAGE=opencloudeu/opencloud-rolling
OC_DOCKER_TAG=5
OC_DOMAIN=cloud.domain.com
CHECK_FOR_UPDATES=true
OC_DATA_DIR=/mnt/storage/data/opencloud/data # this is a CIFS/SMB share
START_ADDITIONAL_SERVICES="antivirus"
COLLABORA_SSL_ENABLE=false
COLLABORA_HOME_MODE=true
COMPOSE_PATH_SEPARATOR=:
RADICALE_DOCKER_IMAGE=opencloudeu/radicale
RADICALE_DOCKER_TAG=latest
RADICALE_DATA_DIR=/mnt/storage/data/opencloud/radicale

OpenCloud Version: 5.2.0 rolling

Additional context

I noticed this issue while running rclone to migrate my data from Nextcloud. At some point, the transfer stopped working with the previously mentioned 401 Unauthorized errors I've been getting.

[micbar]

OC_DATA_DIR=/mnt/storage/data/opencloud/data # this is a CIFS/SMB share

I am not sure if opencloud runs on CIFS/SMB.

[SangeloDev]

OC_DATA_DIR=/mnt/storage/data/opencloud/data # this is a CIFS/SMB share

I am not sure if opencloud runs on CIFS/SMB.

How would I go about figuring out if CIFS is causing the issue here? I'd definitely prefer to keep OC on the CIFS drive, since my host itself has very little storage.