Hdparm

hdparm is a disk utility that allows for low level commands to be sent to the storage device.

Usage

hdparm can be used to on any mass storage device.

Get disk information

Use hdparm -I /dev/disk to show hardware properties of the device.

# hdparm -I /dev/sdb

/dev/sdb:

ATA device, with non-removable media
	Model Number:       ST3000DM001-9YN166                      
	Serial Number:      W1F0SJ27
	Firmware Revision:  CC4B    
	Transport:          Serial, SATA Rev 3.0
Standards:
	Used: unknown (minor revision code 0x0029) 
	Supported: 8 7 6 5 
	Likely used: 8
Configuration:
	Logical		max	current
	cylinders	16383	16383
	heads		16	16
	sectors/track	63	63
	--
	CHS current addressable sectors:    16514064
	LBA    user addressable sectors:   268435455
	LBA48  user addressable sectors:  5860533168
	Logical  Sector size:                   512 bytes
	Physical Sector size:                  4096 bytes
	Logical Sector-0 offset:                  0 bytes
	device size with M = 1024*1024:     2861588 MBytes
	device size with M = 1000*1000:     3000592 MBytes (3000 GB)
	cache/buffer size  = unknown
	Nominal Media Rotation Rate: 7200
Capabilities:
	LBA, IORDY(can be disabled)
	Queue depth: 32
	Standby timer values: spec'd by Standard, no device specific minimum
	R/W multiple sector transfer: Max = 16	Current = ?
	Advanced power management level: 128
	Recommended acoustic management value: 208, current value: 0
	DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 udma5 *udma6 
	     Cycle time: min=120ns recommended=120ns
	PIO: pio0 pio1 pio2 pio3 pio4 
	     Cycle time: no flow control=120ns  IORDY flow control=120ns
Commands/features:
	Enabled	Supported:
	   *	SMART feature set
	    	Security Mode feature set
	   *	Power Management feature set
	   *	Write cache
	   *	Look-ahead
	   *	Host Protected Area feature set
	   *	WRITE_BUFFER command
	   *	READ_BUFFER command
	   *	DOWNLOAD_MICROCODE
	   *	Advanced Power Management feature set
	    	SET_MAX security extension
	   *	48-bit Address feature set
	   *	Device Configuration Overlay feature set
	   *	Mandatory FLUSH_CACHE
	   *	FLUSH_CACHE_EXT
	   *	SMART error logging
	   *	SMART self-test
	   *	General Purpose Logging feature set
	   *	WRITE_{DMA|MULTIPLE}_FUA_EXT
	   *	64-bit World wide name
	    	Write-Read-Verify feature set
	   *	WRITE_UNCORRECTABLE_EXT command
	   *	{READ,WRITE}_DMA_EXT_GPL commands
	   *	Segmented DOWNLOAD_MICROCODE
	   *	Gen1 signaling speed (1.5Gb/s)
	   *	Gen2 signaling speed (3.0Gb/s)
	   *	Gen3 signaling speed (6.0Gb/s)
	   *	Native Command Queueing (NCQ)
	   *	Phy event counters
	   *	READ_LOG_DMA_EXT equivalent to READ_LOG_EXT
	    	DMA Setup Auto-Activate optimization
	    	Device-initiated interface power management
	   *	Software settings preservation
	   *	SMART Command Transport (SCT) feature set
	   *	SCT Write Same (AC2)
	    	unknown 206[7]
	    	unknown 206[12] (vendor specific)
	    	unknown 206[13] (vendor specific)
Security: 
	Master password revision code = 65534
		supported
	not	enabled
	not	locked
	not	frozen
	not	expired: security count
		supported: enhanced erase
	330min for SECURITY ERASE UNIT. 330min for ENHANCED SECURITY ERASE UNIT.
Logical Unit WWN Device Identifier: 5000c500529df4f3
	NAA		: 5
	IEEE OUI	: 000c50
	Unique ID	: 0529df4f3
Checksum: correct

Disk Timing and Benchmark

Use hdparm -t to benchmark how fast the disk can be read. The documentation suggests running this a few times on an idle system for the best result.

Because disks start from the outside of the platter, an offset could be used to test disk performance in another area with the --offset <GiB> parameter. For example, a 8TB disk could be benchmarked with an offset of 4000 GB to see the read speed somewhere near the middle of the platter.

Tuning

You can tune how many sectors the disk is to return for each data transmission to help speed up disk reads. According to the manual, this can help with the OS I/O overhead by 30-50% and may improve throughput between 5%-50%. Eg: hdparm -m16

You can improve sequential reads of large files by setting a hread-ahead value. This specifies how many sectors to read-ahead by in anticipation of them being read later on. Set this with the -a option: hdparm -a16.

Acoustic modes can be set with -M, passing either 128 (for quiet mode), or 254 (fastest).

Tasks

Secure Erase

See: Securely_Erasing_a_Disk#ATA_Secure_Erase

Fix a damaged sector by forcing reallocation

All modern hard drives have a small set of reserved sectors that can be reallocated (or mapped to) by bad sectors as they are detected by the disk firmware. This is usually a transparent process and only happens when the sector is written to. However, because of how filesystems (and possibly the kernel?) functions, it may require some manual intervention with hdparm which we'll go into details below.

Typically, bad sectors will present itself as a device that can't be read or written to at a specific location which results in the filesystem being mounted as read-only or the disk being dropped from a RAID array. Another good clue would be having dd result in an "Input/output error". The Linux kernel would also complain similar to the following (from dmesg):

[Fri Apr  2 20:47:18 2021] sd 2:0:0:0: [sdb] tag#11 FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[Fri Apr  2 20:47:18 2021] sd 2:0:0:0: [sdb] tag#11 Sense Key : Medium Error [current]
[Fri Apr  2 20:47:18 2021] sd 2:0:0:0: [sdb] tag#11 Add. Sense: Unrecovered read error - auto reallocate failed
[Fri Apr  2 20:47:18 2021] sd 2:0:0:0: [sdb] tag#11 CDB: Read(16) 88 00 00 00 00 00 00 00 02 08 00 00 01 f8 00 00
[Fri Apr  2 20:47:18 2021] print_req_error: I/O error, dev sdb, sector 1000 flags 0

Bad sectors can also be found with SMART tests. The LBA of any problematic sectors will be displayed with the smartctl -a output and will look something similar to the following under the self-test section:

SMART Self-test log structure revision number 1
Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
 1  Extended offline    Completed: read failure       40%     12786         4005685760

Knowing which sectors are bad allows us to try making the drive firmware be aware of the issue and allow it to reallocate the bad sector. We can force a low level sector read or write using the hdparm utility. We can try reading the bad sector with hdparm to confirm the issue:

# hdparm --read-sector 4005685760  /dev/sdb
/dev/sdb:
reading sector 4005685760: FAILED: Input/output error

W can try to remap the bad sector by writing to this specific sector with hdparm. The drive may take some time to reallocate. If this fails with an error, something else far worse than bad sectors is probably wrong with the drive. Usually, bad sectors come consecutively and it may help to write to a range of sectors. hdparm doesn't appear to support this, but can be worked around using a loop in bash:

# hdparm --write-sector 4005685760 --yes-i-know-what-i-am-doing /dev/sdb
## This could take some time if the disk needs to reallocate the sector.
/dev/sdb:
re-writing sector 4005685760: succeeded

## In cases where you think there's a range of bad sectors, you can do something like:
# for i in {1000..1010} ; do echo $i ; hdparm --write-sector $i  --yes-i-know-what-i-am-doing  /dev/sdb ; done

If the corrupt sector was reallocated, you should see both the Reallocated_Sector_Ct and Reallocated_Event_Count value increase and the Current_Pending_Sector decrease in the SMART report (ie. from smartctl -a /dev/sdb). Your goal is to try writing to the disk using hdparm until Current_Pending_Sector becomes 0.