Simple script for collectd exec plugin

collectd_plugin_exec_hdd_smart.sh

#!/bin/sh

if [ "$(id -u)" -ne 0 ]; then
  exec sudo "$0"
fi

HOSTNAME="${COLLECTD_HOSTNAME:-`hostname -f`}" # RH / CentOs
#HOSTNAME="${COLLECTD_HOSTNAME:-`hostname`}"     # Debian/Ubuntu
INTERVAL="${COLLECTD_INTERVAL:-60}"
INTERVAL="${INTERVAL%.*}"

while :; do
# this collects at first ata devices
DEVS_TO_SCAN=
DEVS_TO_SCAN="$DEVS_TO_SCAN $(awk '$NF !~ "^(dm-|md|sr|zram|name|sd[a-z]+[1-9])" && $NF != ""  {print "/dev/"$NF}' /proc/partitions)"
#DEVS_TO_SCAN="$DEVS_TO_SCAN /dev/disk/by-id/ata-*"            # ata-$VENDOR-$MODEL-$SERNO
#DEVS_TO_SCAN="$DEVS_TO_SCAN /dev/disk/by-id/wwn-*"            # wwn-$WWN
#DEVS_TO_SCAN="$DEVS_TO_SCAN /dev/disk/by-id/scsi-*"           # scsi-$SCSI($WNN), scsi-$VENDOR-$MODEL-$SERNO
#DEVS_TO_SCAN="$DEVS_TO_SCAN /dev/disk/by-id/usb-*"            # usb-$VENDOR-$MODEL-$SERNO
#DEVS_TO_SCAN="$DEVS_TO_SCAN /dev/sd[a-z] /dev/sda[a-z][a-z]"  # fallback

  found_devs=
  for dev in $DEVS_TO_SCAN; do # to be extended

    # ignore partitions from /dev/disk/by-*/*-part[1-9]*
    case "$dev" in
      *-part[1-9] ) continue
    ;; esac

    # scan devices only once
    realdev="$(readlink -f "$dev")"
    case "$found_devs" in
      *" $realdev "* ) continue ;;
    esac
    found_devs="$found_devs $realdev "

    # skip errors
    smartctl_output="$(smartctl -a "$realdev")"
    if [  "$?" != 0 ]; then
      continue
    fi
    serial="$( echo "$smartctl_output" | sed -nre 's/^Serial Number:[[:space:]]+(.*)$/\1/p' | tr '.-' '__' )"

    instance="${realdev##*/}"
    #instance="$serial"
    instance="${realdev##*/}_${serial}"
    echo "$smartctl_output" | \
      while read attribute _t2 _t3 _t4 _t5 _t6 type _t8 _t9 value _tX; do
        case "$type" in
          Old_age|Pre-fail ) ;; # only use these types
          *) continue ;;
        esac

        case "$attribute" in # only collect certail metrics
          5   ) metric="reallocated_sector-ct"   ;;
          9   ) metric="power_on-hours"          ;;
          12  ) metric="power_cycle-count"       ;;
          168 ) metric="sata_phy_error-count"    ;;
          170 ) metric="bad_block-count"         ;;
          190 ) metric="airflow_temperature-cel" ;;
          194 ) metric="temperature-celsius"     ;;
          197 ) metric="current_pending-sector"  ;;
          198 ) metric="offline-uncorrectable"   ;;
          * ) continue ;;
          esac

          echo "PUTVAL $HOSTNAME/smart-${instance}/$metric interval=$INTERVAL N:$value"
        done
  done
  sleep "$INTERVAL"
done

# collectd plugin config
#LoadPlugin exec
#<Plugin exec>
#  Exec "collectd:disk" "/etc/collectd/collectd.conf.d/plugin_exec_smart.sh"
#</Plugin>

# sudoers
#Cmnd_Alias      COLLECTD_PLUGIN_EXEC_SMART = /etc/collectd/collectd.conf.d/plugin_exec_smart.sh
#collectd        ALL = (root) NOPASSWD: COLLECTD_PLUGIN_EXEC_SMART




# custom types
#raw_read_error          rate:GAUGE:0:U
#spin_up                 time:GAUGE:0:U
#start_stop              count:GAUGE:0:U
#reallocated_sector      ct:GAUGE:0:U
#seek_error              rate:GAUGE:0:U
#seek_time               performance:GAUGE:0:U
#power_on                hours:GAUGE:0:U
#spin_retry              count:GAUGE:0:U
#calibration_retry       count:GAUGE:0:U
#power_cycle             count:GAUGE:0:U
#read_soft_error         rate:GAUGE:0:U
#runtime_bad             block:GAUGE:0:U
#end_to_end              error:GAUGE:0:U
#reported                uncorrect:GAUGE:0:U
#command                 timeout:GAUGE:0:U
#airflow_temperature     cel:GAUGE:0:U
#temperature             celsius:GAUGE:0:U
#hardware_ecc            recovered:GAUGE:0:U
#reallocated_event       count:GAUGE:0:
#current_pending         sector:GAUGE:0:U
#current_pending         sector:GAUGE:0:U
#offline                 uncorrectable:GAUGE:0:U
#udma_crc_error          count:GAUGE:0:U
#multi_zone_error        rate:GAUGE:0:U
#soft_read_error         rate:GAUGE:0:U
#throughput              performance:GAUGE:0:U
#g_sense_error           rate:GAUGE:0:U
#power_off_retract       count:GAUGE:0:U
#load_retry              count:GAUGE:0:U
#load_cycle              count:GAUGE:0:U
#sata_phy_error          count:GAUGE:0:U
#bad_block               count:GAUGE:0:U
#erase                   count:GAUGE:0:U
#bad_block               count:GAUGE:0:U
#bad_cluster_table       count:GAUGE:0:U


# smartctl Attributes
# smartctl -a | awk '$1 ~ "[0-9]+" && $7 ~ "Old_age|Pre-fail" {gsub("-","_", $2); print $1" "tolower($2)" "$10}'
# 1 raw_read_error_rate 176026840
# 3 spin_up_time 0
# 4 start_stop_count 25
# 5 reallocated_sector_ct 0
# 7 seek_error_rate 307256091
# 9 power_on_hours 17826
# 10 spin_retry_count 0
# 12 power_cycle_count 26
# 168 sata_phy_error-count 0
# 184 end_to_end_error 0
# 187 reported_uncorrect 0
# 188 command_timeout 0
# 189 high_fly_writes 175
# 190 airflow_temperature_cel 36
# 191 g_sense_error_rate 0
# 192 power_off_retract_count 21
# 193 load_cycle_count 25
# 194 temperature_celsius 36
# 197 current_pending_sector 0
# 198 offline_uncorrectable 0
# 199 udma_crc_error_count 0

#SSD
# 173 erase-count
# 170 bad_block-count
# 175 bad_cluster_table-count

The following attributes might intesting for SSDs:
Buildin in Rev 3.

 173 Erase_Count 
 170 Bad_Block_Count   
 168 SATA_Phy_Error_Count

TODO:

 175 Bad_Cluster_Table_Count  
 235 Later_Bad_Block
 236 Unstable_Power_Count
 240 Write_Head
 241 Total_LBAs_Written