Hi everyone, I am facing a critical issue with my ZimaOS NAS. It started when my Overseerr container stopped loading posters and metadata, although the Web UI remained accessible.
After a long debugging session, I isolated the root cause. It is not a Docker or Network issue, but a storage issue on the system partition.
The Problem: The root partition (/) is 100% full. Since ZimaOS has a very small system partition (1.2GB), this seems to be preventing the system from creating temporary files required for SSL handshakes, causing all external HTTPS requests to fail.
Diagnosis & Logs:
-
Symptoms: Running
curl -v https://www.google.comfails instantly with:OpenSSL SSL_connect: SSL_ERROR_SYSCALLHowever,curl http://www.google.com(HTTP) works perfectly. -
Disk Space: The command
df -h /shows:/dev/root 1.2G 1.2G 0 100% / -
Deep Analysis: We inspected the filesystem using a bind mount to look “under” the active overlays (
sudo mount --bind / /tmp/root_debug). We found that/usris reporting a size of 2.8GB inside the bind mount. This suggests that the writable layer of the OverlayFS is completely saturated.
What I have tried so far:
-
Ran
docker system prune(it only freed about 40MB). -
Cleared logs in
/var/logand vacuumed the systemd journal. -
Tried removing unused firmware in
/usr/lib/firmware.
Despite these efforts, df -h / remains stuck at 100%. I suspect there might be “masked” files (e.g., data written to /var or /usr before the external Data disk was mounted) or zombie files that the system is not releasing.
My Question: On ZimaOS (which uses a read-only/overlay filesystem structure), what is the correct procedure to identify and clean the writable upper layer to unblock the root partition without performing a full factory reset?
Thanks in advance for any help!