$ guides / smartctl-disk-monitoring / smartctl-seek-error-rate ▌

Operations Guides

Seek_Error_Rate: head-positioning wear and the Seagate raw-value trap

A monitoring dashboard shows Seek_Error_Rate with a raw value of 200,009,354,607 on a Seagate drive. The on-call engineer pages the storage team. The replacement drive goes into the same bay and shows the same number. This cycle repeats across fleets because most monitoring tools and most operators do not know how Seagate encodes this attribute.

Seek_Error_Rate (ATA attribute ID 7) tracks the accuracy of the HDD actuator arm as it positions read/write heads over target tracks. On most non-Seagate drives, the raw value is a straightforward error count or rate. On Seagate drives, the raw value is a packed composite encoding both total seek operations and seek errors, producing numbers in the billions on perfectly healthy hardware. Alerting on the raw value guarantees false positives on every Seagate drive in the fleet.

The signal that matters is the normalized VALUE column, not the RAW_VALUE column. A declining normalized value indicates actuator or head-positioning degradation. A stable normalized value, regardless of how large the raw value looks, means the positioning system is operating within spec.

What this means

Seek_Error_Rate reports how often the drive’s actuator arm fails to land the read/write head on the target track. Each miss triggers a retry: the arm repositions and attempts the read or write again. Occasional retries are normal. A rising rate of retries means the mechanical positioning system is losing precision, which increases latency and, in severe cases, risks head crashes or data corruption.

The attribute is HDD-only. SSDs have no actuator arm, no platters, and no seek mechanism. If an SSD exposes attribute ID 7, it is a repurposed field with different semantics or a vestigial placeholder. Do not apply HDD interpretation logic to SSD attribute tables.

The Seagate encoding trap

Seagate packs two values into the 48-bit RAW_VALUE field for attribute ID 7 (and for ID 1, Raw_Read_Error_Rate). The structure is:

Lower bits: total number of seek operations performed since the counter last reset
Upper bits: number of seek errors during that same window

A raw value of 200,009,354,607 might represent billions of seeks with fewer than a hundred errors. The raw value is not an error count. It is a composite that will always look enormous on a drive that has been operating for any length of time.

The normalized VALUE is computed from the ratio of total seeks to seek errors. On Seagate drives, the scale is logarithmic and inverted: a higher normalized value means fewer errors relative to total seeks. Approximate reference points:

Normalized VALUE	Approximate error rate
60 to 100	1 error per million seeks or better (healthy)
50	roughly 10 errors per million seeks
40	roughly 100 errors per million seeks (concerning)
Near THRESH	Vendor-defined failure boundary

Seagate’s counters roll over when the total seek count reaches its bit-width limit. At that point, the normalized VALUE may jump upward (appear to improve). A rising normalized value on a Seagate drive does not mean the drive healed. It means the internal counters reset and the ratio is being recalculated from scratch.

Non-Seagate drives

Western Digital, Toshiba, and HGST typically use a simpler encoding for ID 7. The raw value is often a direct error count or a small composite. A raw value of 0 is expected on a healthy drive. Any nonzero raw value may warrant investigation, but always confirm against the normalized VALUE trend before taking action.

Common causes

Cause	What it looks like	First thing to check
Seagate raw-value encoding	Raw value in the billions, normalized VALUE stable at 60 or above	Confirm drive vendor with `smartctl -i`. Check normalized VALUE, not RAW_VALUE.
Environmental vibration	Normalized VALUE declining on multiple drives in the same chassis or bay	Check G-Sense Error Rate (ID 221) for the same drives
Actuator or head wear	Normalized VALUE declining on a single drive, G-Sense stable	Check Reallocated Sector Count (ID 5) and Spin Retry Count (ID 10)
Thermal expansion	Normalized VALUE dips correlate with temperature spikes above 50C	Check Temperature_Celsius (ID 194) trend
Monitoring tool misconfiguration	Tool reports FAILED or critical on a healthy Seagate drive	Check whether the tool passes vendor-specific smartctl flags for Seagate encoding

Quick checks

All commands below are read-only and safe to run on production drives.

# Identify the drive vendor and model
smartctl -i /dev/sdX | grep -iE "Model|Vendor|Firmware"

# Check Seek_Error_Rate attribute (ID 7)
smartctl -A /dev/sdX | grep -E "^  7"

# For Seagate drives: the raw value is a composite.
# To attempt to decode the error count from the upper 16 bits:
# <!-- TODO: verify smartctl -v format string for Seagate ID 7 decoding.
# raw48:54 may not be a valid smartmontools format string.
# The standard smartctl -v options (raw16(raw16), raw16(raw24), etc.)
# do not cleanly split the Seagate 48-bit composite into 16+32.
# Manual bit extraction from the raw value may be required. -->
smartctl -A -v 7,raw48:54 /dev/sdX | grep -E "^  7"

# Check vibration indicator (G-Sense Error Rate, ID 221)
smartctl -A /dev/sdX | grep -E "^221"

# Check spin retry count (ID 10) for mechanical correlation
smartctl -A /dev/sdX | grep -E "^ 10"

# Check reallocated sectors (ID 5) for media degradation
smartctl -A /dev/sdX | grep -E "^  5"

# Check temperature (ID 194) for thermal correlation
smartctl -A /dev/sdX | grep -E "^194"

# Check raw read error rate (ID 1) for broader mechanical pattern
smartctl -A /dev/sdX | grep -E "^  1"

# Full attribute table for comprehensive review
smartctl -A /dev/sdX

How to diagnose it

The diagnostic flow depends entirely on the drive vendor and on whether the normalized VALUE is actually declining. The raw value alone is diagnostic noise on Seagate drives.

flowchart TD
    A["Seek_Error_Rate raw value looks alarming"] --> B{"Seagate drive?"}
    B -->|Yes| C["Raw value is a packed composite.
It is NOT an error count"]
    B -->|No| D["Check raw value against
vendor baseline"]
    C --> E{"Normalized VALUE
trending down?"}
    D --> E
    E -->|No| F["Positioning system is healthy.
Fix monitoring if it alerted"]
    E -->|Yes| G{"G-Sense Rate ID 221
also climbing?"}
    G -->|Yes| H["Environmental vibration.
Fix: mechanical damping"]
    G -->|No| I{"Reallocated ID 5 or
Spin Retry ID 10 rising?"}
    I -->|Yes| J["Active degradation.
Plan drive replacement"]
    I -->|No| K["Early actuator wear.
Monitor, plan replacement"]

Identify the vendor. Run smartctl -i /dev/sdX and check the Model or Family line. If it is a Seagate drive, the raw value is a composite. Stop reading the RAW_VALUE column for this attribute.
Check the normalized VALUE trend. A single snapshot tells you almost nothing. You need at least two data points separated by days or weeks. Compare the VALUE column against the THRESH column. If VALUE is well above THRESH and stable, the drive is fine. If VALUE is declining toward THRESH, the positioning system is degrading.
Decode Seagate raw values if you need the detail. For Seagate drives, the upper 16 bits of the 48-bit raw value represent the error count and the lower 32 bits represent total seek operations. Manual bit extraction may be required if smartctl cannot natively split the field. Compare the error count against the total operation count to understand the actual error ratio.
Check for vibration. Look at G-Sense Error Rate (ID 221). If it is climbing alongside Seek_Error_Rate, the root cause is physical vibration, not a drive defect. Vibration from adjacent spinning drives in dense disk shelves, fans, HVAC, or server relocation can all cause seek errors. The fix is mechanical, not a drive swap.
Check for correlated mechanical failure. Look at Spin Retry Count (ID 10) and Reallocated Sector Count (ID 5). If seek errors are rising alongside spin retries, the drive’s mechanical systems are failing together. If reallocated sectors are also growing, the seek errors may already be causing data corruption as heads land on wrong tracks or write to incorrect locations.
Check temperature correlation. If the normalized VALUE dips during periods when Temperature_Celsius (ID 194) exceeds 50C, thermal expansion of the platters may be affecting head positioning accuracy. The fix is cooling, not drive replacement.

Metrics and signals to monitor

Signal	Why it matters	Warning sign
Seek_Error_Rate (ID 7) normalized VALUE	Direct indicator of head-positioning accuracy	Declining trend over days or weeks
G-Sense Error Rate (ID 221)	Distinguishes vibration-induced seek errors from mechanical wear	Any growth, especially on multiple drives in the same chassis
Spin Retry Count (ID 10)	Correlated mechanical failure of spindle motor or power delivery	Any nonzero value
Reallocated Sector Count (ID 5)	Seek errors causing data corruption that triggers remapping	Any growth from baseline
Raw Read Error Rate (ID 1) normalized VALUE	Broader pattern of mechanical degradation across the read path	Declining trend alongside ID 7
Temperature_Celsius (ID 194)	Thermal expansion affecting positioning precision	Spikes above 50C correlating with seek error increases

Fixes

Fix your monitoring first

If your monitoring tool alerts on the raw value of ID 7 for Seagate drives, the tool is broken, not the drive. Configure it to track the normalized VALUE column instead. If the tool does not support per-attribute normalization or per-vendor interpretation, file a bug or configure per-drive smartctl arguments that handle the Seagate composite encoding.

Tools that do not account for Seagate encoding will show false failures on every Seagate HDD. The result is alert fatigue: operators learn to ignore SMART alerts and miss real failures on other drives.

Vibration-induced seek errors

If G-Sense Error Rate confirms vibration as the root cause, the fix is mechanical:

Rubber mounting grommets: Install anti-vibration mounts if the chassis supports them.
Chassis stiffening: A flexing chassis transmits vibration between drives. Ensure the drive bay is rigid and properly torqued.
Stagger spin-up: If all drives spin up simultaneously after a power cycle, the aggregate vibration and inrush current can cause seek errors and spin retries across the shelf. Enable staggered spin-up in the HBA or controller BIOS.
Relocate the drive: If one bay is particularly noisy (adjacent to a fan or power supply), move the drive to a quieter position.

Do not replace the drive. The replacement will exhibit the same symptoms in the same environment.

Actuator or head wear

If the normalized VALUE is declining without a vibration or thermal correlation, the actuator assembly is wearing. This is progressive and non-recoverable.

Begin replacement planning. The drive will continue to degrade. Do not wait for SMART overall health to change to FAILED. It may not happen until the drive is nearly dead.
Check for data impact. If reallocated sectors are growing alongside seek errors, the positioning errors are already causing media damage. Prioritize data migration.
Run an extended self-test: smartctl -t long /dev/sdX. If the test reports a servo or seek failure, the drive has confirmed mechanical damage and should be replaced immediately.

Thermal-induced seek errors

If temperature correlation is confirmed, address the cooling problem. Check chassis fans, ambient temperature, and airflow obstruction. The drive itself may be healthy once temperatures return to the normal operating range (25 to 45C for HDDs). Monitor the normalized VALUE after the thermal fix to confirm it stabilizes.

Prevention

Baseline every drive at deployment. Capture the initial normalized VALUE for ID 7 (and ID 1) so you have a reference point for trend analysis. Without a baseline, you cannot distinguish a declining trend from a stable low value.
Monitor the normalized VALUE trend, not the raw value. This applies to all Seagate attributes that use packed encoding (IDs 1, 7, and 195).
Alert on rate of change, not absolute value. A drive that has held normalized VALUE at 60 for three years is healthy. A drive that dropped from 80 to 60 in two weeks is degrading.
Schedule periodic extended self-tests. These surface servo and seek failures before production I/O hits them. Weekly short tests and monthly long tests are standard practice.
Track G-Sense Error Rate fleet-wide. If multiple drives in the same chassis show vibration events, the problem is environmental. Fix the chassis, not the drives.

How Netdata helps

Netdata collects SMART attributes per drive at per-second resolution, including the normalized VALUE and raw VALUE for ID 7. Trending the normalized VALUE over time surfaces the actual degradation signal without the Seagate raw-value noise.
Anomaly detection on the normalized VALUE series flags a declining trend automatically, without manual static thresholds that may not apply across mixed vendor fleets.
Correlating Seek_Error_Rate with G-Sense Error Rate (ID 221), Spin Retry Count (ID 10), and Reallocated Sector Count (ID 5) on the same dashboard lets you distinguish vibration from mechanical wear without switching between terminal sessions.
Temperature overlay lets you check whether seek error increases correlate with thermal events, which changes the response from drive replacement to cooling remediation.
Fleet-level views show whether seek error degradation is isolated to one drive or affecting multiple drives in the same chassis, which is the signature of an environmental problem rather than individual drive failure.