KingDezigns — SSD Hardening Summary

Objective

Harden both NAS08 and NAS16 to extend SSD lifespan by ensuring TRIM is running correctly, noatime is active, and SMART health is being monitored — with all scheduled jobs visible in OMV's scheduler for easy management.

NAS08 — 5 drives (1x PNY CS900 + 4x Crucial BX500) NAS16 — 1 drive (unknown model, standard SMART attrs)

Investigation steps — both units

1 Verify TRIM support on all drives

lsblk --discard /dev/sda

Confirm DISC-GRAN and DISC-MAX are non-zero before enabling TRIM.

Both units: DISC-GRAN 4K / DISC-MAX 4G — TRIM fully supported on all drives.

2 Check if systemd fstrim.timer was already active

systemctl status fstrim.timer

Determine whether a TRIM schedule already existed before creating a new one.

Both units: timer active and enabled, running weekly. NAS08 next run Monday 12:33 AM — NAS16 next run Monday 1:37 AM.

3 Confirm which volumes were being trimmed

journalctl -u fstrim.service --no-pager

Verify the timer was actually running and catching all mounted volumes.

Both units: / and /boot/firmware trimmed successfully on May 11 and May 18. NAS08 ~400 GiB freed per run. NAS16 ~228 GiB on first run, 3.3 GiB on second (less free space).

4 Check noatime mount status

findmnt -t ext4,xfs,btrfs -o TARGET,SOURCE,OPTIONS

Confirm noatime was already set on all filesystems to prevent reads generating writes.

Both units: every mount point already showing noatime — handled automatically by the openmediavault-flashmemory plugin. No action required.

5 Inspect SMART attributes per drive

sudo smartctl -A /dev/sda
sudo smartctl -A /dev/sdb   # NAS08 data drives

Identify vendor-specific SMART attribute names before writing the monitoring script — attribute names differ significantly between PNY and Crucial.

NAS08 PNY CS900: uses SSD_Life_Left (attr 231), Bad_Blk_Ct_Lat/Erl (attr 170), no pending/uncorrectable attrs. NAS08 Crucial BX500: uses Percent_Lifetime_Remain (attr 202), Reallocate_NAND_Blk_Cnt (attr 5), Reported_Uncorrect. NAS16 drive: uses standard attribute names — no vendor mapping required.

What was already handled by OMV

noatime: Already active on all filesystems on both units via the openmediavault-flashmemory plugin. No manual configuration required.

Swap disabled, logs in tmpfs: Also handled automatically by the flashmemory plugin — swap off, /var/log and /tmp in RAM, reducing write amplification significantly.

⚠ Note

The systemd fstrim.timer was already running weekly on both units but was invisible in OMV's scheduler. It was replaced with an OMV Scheduled Job so all jobs are visible and manageable in one place.

TRIM report script — created and deployed

A custom script was written to replace the bare fstrim call with a full monitoring run that emails an HTML report matching the KingDezigns backup monitor style (dark header, badge/banner, section rows, cards, footer).

The email report includes: TRIM results per volume with total bytes freed, SSD SMART health per drive (health, temperature, wear, power-on hours, sector counts), filesystem usage with colour-coded warnings at 75% and 90%, and system info (OMV version, kernel, IP, uptime, load, memory, next run time).

Script location

/usr/scripts/omv/fstrim-report.sh

Log location

/var/log/fstrim/fstrim-YYYY-MM-DD.log   (90-day retention)

Deploy (preferred method — sudo nano)

sudo nano /usr/scripts/omv/fstrim-report.sh
# paste script content, save with Ctrl+O, exit with Ctrl+X
sudo chmod +x /usr/scripts/omv/fstrim-report.sh

Test run

sudo systemctl start fstrim.service
sudo journalctl -u fstrim.service -f

Email delivery

sendmail -t   (via OMV's configured SMTP relay)

⚠ SMART attribute fixes required

Initial script used generic attribute names. Two rounds of fixes were needed after testing:

Round 1 — Crucial BX500 fix:
  Reallocated: Reallocated_Sector → Reallocate_NAND_Blk_Cnt
  Pending:     Current_Pending_Sector → Current_Pending_ECC_Cnt
  Uncorrect:   Offline_Uncorrectable → Reported_Uncorrect
  Temperature: $10 → $10+0  (strips "32 (Min/Max 25/56)" format)

Round 2 — PNY CS900 fix:
  Wear:        Wear_Leveling_Count → SSD_Life_Left
  Reallocated: Reallocate_NAND_Blk_Cnt → Bad_Blk_Ct_Lat/Erl (first value before /)

Each pattern now tries vendor-specific names first, then falls back to generic names — script works correctly on any brand.

Migrating from systemd timer to OMV scheduler

The systemd fstrim.timer was disabled on both units and replaced with OMV Scheduled Jobs so all scheduled tasks are visible in one place.

1 Disable systemd timer (both units)

sudo systemctl disable fstrim.timer
sudo systemctl stop fstrim.timer

Prevent the systemd timer from running alongside the new OMV job — avoids double execution.

Timer stopped and disabled on both NAS08 and NAS16.

2 Remove systemd override (both units)

sudo rm /etc/systemd/system/fstrim.service.d/override.conf
sudo systemctl daemon-reload

Clean up the override that pointed fstrim.service to the script — no longer needed since OMV cron calls the script directly.

Override removed, systemd reloaded cleanly on both units.

3 Add OMV Scheduled Job (each unit)

System → Scheduled Jobs → Add
Command: /usr/scripts/omv/fstrim-report.sh
User:    root
NAS08:   Wednesday at 1:00 AM
NAS16:   Wednesday at 1:30 AM

Wednesday was chosen as it sits midweek, away from Sunday ZFS scrubs and the rolling every-3-day backup jobs.

Both units now show TRIM Report in OMV scheduler alongside all other scheduled jobs.

Final OMV scheduled jobs — both units

NAS08

Schedule	Script	Tag
Every 3 days at 2:00 AM	`/usr/scripts/omv/nas08-backup.sh`	NAS08 Backup
Daily at 12:00 AM	`/usr/scripts/zfs/nas08_zfs_report.sh`	ZFS Status Report
Sunday at 3:00 AM	`/usr/scripts/zfs/nas08_zfs_scrub.sh`	ZFS Scrub
Wednesday at 1:00 AM	`/usr/scripts/omv/fstrim-report.sh`	TRIM Report

NAS16

Schedule	Script	Tag
Every 3 days at 2:00 AM	`/usr/scripts/omv/nas16-backup.sh`	NAS16 Backup
Every 2 days at 2:30 AM	`bash /usr/scripts/nas/nas08_to_nas16_sync.sh`	NAS Drive Backup
Daily at 12:00 AM	`/usr/scripts/zfs/nas16_zfs_report.sh`	ZFS Status Report
Sunday at 3:00 AM	`/usr/scripts/zfs/nas16_zfs_scrub.sh`	ZFS Scrub
Every 15 minutes	`/usr/scripts/ddns/ddns_update.sh`	DDNS Update
Wednesday at 1:30 AM	`/usr/scripts/omv/fstrim-report.sh`	TRIM Report

Lessons learned

Lesson	Detail
Always investigate before configuring	Both noatime and fstrim were already handled by OMV's flashmemory plugin — checking first avoided duplicating work.
SMART attribute names are vendor-specific	PNY and Crucial use completely different attribute names for wear, reallocated sectors, and temperature format. Always run `smartctl -A /dev/sdX` first to map the real names.
systemd timers are invisible in OMV scheduler	The default fstrim.timer runs correctly but doesn't appear in OMV's job list — migrate to OMV Scheduled Jobs for full visibility alongside ZFS scrubs and backups.
Never run script without sudo for testing	Running fstrim-report.sh without sudo causes permission denied on /var/log/fstrim/ and returns N/A for all SMART data — always test with sudo or via systemctl start.
Use sudo nano over cp for file edits	Preferred method for placing and editing scripts on the NAS — faster navigation to the correct path, avoids confusion about source file location.
Inline TRIM (discard mount option) is not recommended	Real-time per-delete TRIM adds latency under heavy load. Periodic batch TRIM via fstrim is the correct approach for NAS workloads.
PNY CS900 has no pending/uncorrectable attributes	N/A for those fields is correct and expected — not a script bug. The drive simply does not expose those attributes.

Documentation updated

File	What changed
`server_nas08.md`	Added drive inventory table, full SSD Hardening & Maintenance section (TRIM, noatime, flashmemory plugin, SMART attribute mapping), and complete OMV Scheduled Jobs table.
`server_nas16.md`	Added drive inventory table with Power_Cycle_Count monitoring note, full SSD Hardening & Maintenance section, and complete OMV Scheduled Jobs table including NAS drive backup sync job.