Interpreting SMART Attribute Values

Interpreting the output from smartctl is tricky because some of the parameters represent packed values that have to be processed first. For example, Seagate describes its SMART parameters in their hard-drive SMART attribute documentation that lists all the parameters and how the values are packed.

For example, reading Raw_Read_Error_Rate or Seek_Error_Rate seems pretty alarming due to the high values but the value has to be converted into hexadecimal and then broken up into various sub-values in order to obtain meaningful readings.

The following script follows the Seagate attribute manual and computes the real number of failures. Alternatively, for graphical interfaces, gsmartctrl is a smartctl GNOME GUI that can also highlight suspicious or bad values while accounting for the values that look alarming.

#!/bin/sh
 
#                                                                                #
# ███████╗██╗    ██╗███████╗███████╗████████╗ ██████╗  ██████╗  ██████╗ ██████╗  #
# ██╔════╝██║    ██║██╔════╝██╔════╝╚══██╔══╝██╔════╝ ██╔═══██╗██╔═══██╗██╔══██╗ #
# ███████╗██║ █╗ ██║█████╗  █████╗     ██║   ██║  ███╗██║   ██║██║   ██║██║  ██║ #
# ╚════██║██║███╗██║██╔══╝  ██╔══╝     ██║   ██║   ██║██║   ██║██║   ██║██║  ██║ #
# ███████║╚███╔███╔╝███████╗███████╗   ██║   ╚██████╔╝╚██████╔╝╚██████╔╝██████╔╝ #
# ╚══════╝ ╚══╝╚══╝ ╚══════╝╚══════╝   ╚═╝    ╚═════╝  ╚═════╝  ╚═════╝ ╚═════╝  #
#                                                                                #
#                                      IT-Beratung mit Fokus auf Datensicherheit #
#                                                                                #
#                            www.sweetgood.de                                    #
#                                                                                #
# Copyright        : All rights reserved!
# Repository url   : https://codeberg.org/SWEETGOOD/shell-scripts/
# Author           : SWEETGOOD
# Filename         : parse-raw-smart-values-seagate.sh
# Created at       : 13.05.2024
# Last changed at  : 13.05.2024
# Version          : 1.0
# License          : CC BY-SA 4.0 Deed
# Description      : Checks RAW encoded prefailure values on SEAGATE drives
#                    Inspiration from: https://www.disktuna.com/big-scary-raw-s-m-a-r-t-values-arent-always-bad-news
# Requirements     : smartctl bc
 
# If no parameter was given
if [ $# = 0 ]
then
    exit 255
fi
 
LC_ALL=de_DE.UTF-8
 
# Function to split a number into two 32-bit integers (created using https://codingfleet.com/code-converter/javascript/bash/)
split() {
    # Convert the number to binary string
    n=$(echo "obase=2; $1" | bc)
 
    # Pad with zeros to make it 64-bit
    while [ ${#n} -lt 64 ]; do
        n="0$n"
    done
 
    # Split the binary string into two 32-bit parts
    lo=$(echo "${n}" | cut -c1-32)
    hi=$(echo "${n}" | cut -c33-64)
 
    echo "$lo $hi"
}
 
# Check device model
DEVICE=$(/usr/sbin/smartctl -A -i /dev/"$1" | awk -F ' ' '/Device Model/{print $3}')
if [ "$(echo "${DEVICE}" | cut -c1-2)" != "ST" ]; then
        echo "No Seagate drive. Exiting."
        exit
fi
 
printf '%sn\n' "Device Model: ${DEVICE}"
printf '%sn\n' "Serial Number: $(/usr/sbin/smartctl -A -i /dev/"$1" | awk -F ' ' '/Serial Number/{print $3}')"
 
RRER=$(split "$(/usr/sbin/smartctl -A -i /dev/"$1" | awk -F ' ' '/Raw_Read_Error_Rate/{print $10}')")
SER=$(split "$(/usr/sbin/smartctl -A -i /dev/"$1" | awk -F ' ' '/Seek_Error_Rate/{print $10}')")
HECCR=$(split "$(/usr/sbin/smartctl -A -i /dev/"$1" | awk -F ' ' '/Hardware_ECC_Recovered/{print $10}')")
 
RRER1=$(echo "${RRER}" | awk '{ print $1; }')
RRER2=$(echo "${RRER}" | awk '{ print $2; }')
echo "Raw_Read_Error_Rate: $(echo "ibase=2;obase=A;${RRER1}" | bc) errors in $(numfmt --grouping "$(echo "ibase=2;obase=A;${RRER2}" | bc)") operations."
 
SER1=$(echo "${SER}" | awk '{ print $1; }')
SER2=$(echo "${SER}" | awk '{ print $2; }')
echo "Seek_Error_Rate: $(echo "ibase=2;obase=A;${SER1}" | bc) errors in $(numfmt --grouping "$(echo "ibase=2;obase=A;${SER2}" | bc)") operations."
 
HECCR1=$(echo "${HECCR}" | awk '{ print $1; }')
HECCR2=$(echo "${HECCR}" | awk '{ print $2; }')
echo "Hardware_ECC_Recovered: $(echo "ibase=2;obase=A;${HECCR1}" | bc) errors in $(numfmt --grouping "$(echo "ibase=2;obase=A;${HECCR2}" | bc)") operations."
 
# If there is any error
if [ "$(echo "ibase=2;obase=A;${RRER1}" | bc)" != 0 ] || [ "$(echo "ibase=2;obase=A;${SER1}" | bc)" != 0 ] || [ "$(echo "ibase=2;obase=A;${HECCR1}" | bc)" != 0 ]; then
        exit 1
fi

Some parameters seem universal to the hard-drive brand and it is a good idea to monitor them in order to be able to tell whether the drive is falling.

• SMART 5: Reallocated_Sector_Count, 1-4 keep an eye on it, more than 4 replace
• SMART 187: Reported_Uncorrect, 1 or more replace
• SMART 188: Command_Timeout, 1-13 keep an eye on it, more than 13 replace
• SMART 197: Current_Pending_Sector_Count, 1 or more replace
• SMART 198: Offline_Uncorrectable, 1 or more replace

And for SSDs:

• SMART 5: Reallocate_NAND_Blk_Cnt, these are blocks that cannot be reprogrammed anymore; too many of these and the drive is ageing,
• SMART 202: Percent_Lifetime_Remain, to be read as such that when this value reaches zero, then the drive is dead