Umbrel 1.7 shows persistent “Storage problem detected” screen after reboot on Raspberry Pi 5 even after NVMe is mounted successfully

## Bug summary After updating to **Umbrel version 1.7** and rebooting the system, Umbrel displayed the **"Storage problem detected"** screen. After checking whether the SSD/NVMe was connected, the drive was later detected and mounted successfully again, and Umbrel services came back up. However, the UI still continued showing the same storage error screen. This makes it look like the system remains stuck in a storage failure state in the UI even after the storage backend has recovered. ## Environment - Device: Raspberry Pi 5 - Storage: NVMe - Umbrel version: 1.7 - Access method: `umbrel.local` / local IP ## What happened 1. Updated Umbrel to **version 1.7** 2. Rebooted the system 3. After reboot, Umbrel showed the **"Storage problem detected"** screen 4. I checked whether the SSD/NVMe was properly connected 5. After that, the drive was detected and mounted successfully 6. Even after recovery, the UI still continued showing the same storage problem screen ## Expected behavior After the NVMe is detected again, filesystem check passes, external storage mounts successfully, and Umbrel services are running, the UI should leave the storage error screen and return to normal operation. ## Actual behavior The UI still shows the storage problem screen even though: - the NVMe is detected, - the partition is mounted, - `umbrel-external-storage.service` completes successfully, - `umbrel.service` is active, - and multiple app containers are running. ## Evidence collected ### NVMe detected and mounted `lsblk` shows: - `nvme0n1` - model: `KINGSTON SNV3S1000G` - partition `nvme0n1p1` - mounted at `/mnt/root/mnt/data` ### External storage service succeeded `umbrel-external-storage.service` reports: - device found - ext4 detected - filesystem check clean - Umbrel install detected - bind mounts completed - `Mount script completed successfully!` ### Umbrel daemon is running `umbrel.service` is `active (running)`. ### Containers are up Multiple containers are running after boot, including core/auth and several apps. ### Relevant server-side errors `journalctl -u umbrel.service -b` shows repeated authentication/token-related errors such as: - `Missing token` - `Invalid token` - `jwt must be provided` These errors appear during requests to endpoints such as: - `/hardware.raid.checkRaidMountFailureDevices` - `/hardware.internalStorage.getDevices` ## Possible root cause This may be a stale frontend/system state after a temporary storage failure that was later recovered, or a token/session issue causing the UI to keep interpreting the system as failed. I cannot confirm the exact root cause, but the storage backend appears to have recovered while the UI still shows the failure screen. ## Steps observed 1. Update to Umbrel 1.7 2. Reboot the Raspberry Pi 5 3. UI shows the storage problem screen 4. Check whether the NVMe is properly connected 5. System later detects and mounts the NVMe successfully 6. UI still remains on the storage problem screen ## Impact The system appears operational at the service/container level, but the UI remains blocked by an error screen indicating storage failure. ## Additional notes - The storage issue first appeared immediately after the update to version 1.7 and reboot. - The screenshot attached is the exact screen shown after reboot. - Manual restart of `umbrel-external-storage.service` after successful mount caused secondary errors related to already-mounted filesystem/swap handling, so the main issue seems to be the persistent UI state after recovery, not the drive remaining absent. - Full logs and screenshots can be provided if needed. ## Screenshot <img width="1067" height="761" alt="Image" src="https://github.com/user-attachments/assets/75153010-31dd-4e28-9430-1b866524f476" />