A storage pool becoming read-only is a protection symptom, not a diagnosis. In this community case, the user reseated the NVMe drives and rebooted, and the storage returned. That suggests a connection or transient device issue was plausible, but the thread did not prove the update itself caused the problem.
Read-Only Can Be a Protective Response to Storage Errors
Linux filesystems and storage stacks can restrict writes when they encounter serious I/O or consistency problems. On Btrfs, the Btrfs scrub behavior documents checksum verification and repair behavior for supported replicated profiles, but scrub is only one tool and does not replace device-health investigation.
Check Device Health and Connection Stability
The community reply suggested SMART health, system logs, NVMe seating, power, and backplane/cable stability. The smartctl drive diagnostics provides the drive-health diagnostic layer for supported devices.
The Zima PCIe ecosystem is relevant when NVMe devices are attached through PCIe adapters or expansion hardware. A flaky link can surface as a filesystem problem even when the filesystem itself is not the original cause.
Do Not Format the Array as the First Fix
One reply described rebuilding their own RAID after a similar event, but that is not a safe first response. If the data matters, use the RAID recovery workflow recovery principles and verify an independent backup before recreating arrays or filesystems.
The ZimaOS backup workflow is the right next step once the storage is readable again. If the problem recurs, capture logs while the failure exists rather than relying only on the clean state after a reboot.
A Practical Checklist If It Happens Again
- Stop unnecessary writes and do not initialize or format anything.
- Check the Storage/RAID health state.
- Review kernel and system logs for I/O, NVMe, filesystem, or RAID errors.
- Check drive health where SMART/NVMe diagnostics are available.
- Inspect physical seating, adapters, power, and cooling.
- Back up important data as soon as the filesystem is safely readable.
Bottom Line
Reseating the NVMe drives restored this user's storage, but that should be treated as evidence of a possible transient hardware/path issue rather than a permanent repair. A recurring read-only state deserves drive, RAID, filesystem, log, and connection diagnostics before any formatting or rebuild.
