URGENT: TS-253D QTS 5.2.9 — Both HDDs healthy, RAID1 superblocks clean, storage pool missing, raid.conf empty after NVMe HAL configuration change

English Storage
, , , ,
1

We are investigating a storage-pool outage on a QNAP TS-253D running QTS 5.2.9.3499.

Hardware

  • 2 × 8 TB Seagate HDDs configured as RAID 1

  • 1 × 1 TB FIKWOT FX991 NVMe SSD connected via PCIe

  • Both HDDs are detected by QTS and report healthy SMART status

  • Original data volume and iSCSI LUN are inaccessible

Incident background

We were investigating why QTS would not allow the NVMe SSD to be used for a storage pool.

During troubleshooting, /etc/hal.conf was backed up and the [Enclosure_23] protocol value was changed from 5 to 9. This was subsequently reverted to its original value.

The storage pool is now missing from Storage & Snapshots. We have not established whether the storage configuration failed before or after the HAL modification.

Confirmed diagnostics

Both HDD data partitions retain matching RAID1 superblocks:

  • Devices: /dev/sda3, /dev/sdb3

  • Array UUID: 777f0148:20914425:ea5cbfef:8f01dbeb

  • RAID level: RAID1

  • State: clean

  • Events: 274358 on both disks

  • Device roles: 0 and 1

  • Array state: AA

  • Last superblock update: October 9, 2026, 10:21:01

However:

  • /dev/md1 is absent from /proc/mdstat

  • pvs, vgs, and lvs show no active data volume group

  • The active /etc/config/raid.conf contains only [Global] and raidBitmap = 0x0

  • An available raid.conf.bkup references /dev/md2 and a different UUID, apparently associated with the NVMe configuration

  • Storage Manager reports disk/pool mapping errors

  • The existing iSCSI target and LUN names remain visible, but LUN capacity is unavailable

Questions for experienced QNAP users and QNAP Support

  1. What could cause QTS to lose the RAID1 registration while both disks retain matching clean superblocks?

  2. Which logs or configuration databases record when RAID entries are removed from raid.conf?

  3. Can modifying the NVMe enclosure protocol in hal.conf cause HAL or Storage Manager to rewrite RAID registration?

  4. Where does QTS 5.2.9 retain automatic backups of raid.conf, LVM configuration, and storage-pool metadata?

  5. Is the supported Attach and Recover Storage Pool workflow appropriate when the original HDD RAID is not assembled?

  6. What is the safest method to recover the original pool without overwriting metadata or rebuilding the array?

2

Was the SSD configured as a cache?
The SSD manufacturer is not certified for QNAP.

3

Who changed this value? Did you do this? I would suggest NEVER EVER NEVER EVER manually modifying any files that have to do with storage.

At this point, I would recommend opening a ticket with QNAP and let the support team help recover things.

4

Please open a support ticket as soon as possible so our Support Team can help recover the storage pool. We’d also recommend not modifying any system configuration files (such as .conf files) yourself, as this may make recovery more difficult. Thanks!

QNAP Helpdesk