Email reporting for disk health (SMART)

smart_report.sh

#!/bin/sh
 
### Parameters ###
smartctl="/usr/sbin/smartctl"
logfile="/tmp/smart_report.tmp"
email="your_email@your_domain.com"
subject="SMART Status Report for `hostname -f`"
drives="sda sdb"
tempWarn=40
tempCrit=45
sectorsCrit=10
warnSymbol="?"
critSymbol="!"
 
### Set email body ###
echo "<h2>$subject</h2>"
echo "<pre style=\"font-size:14px\">" >> $logfile
 
###### summary ######
(
    echo ""
    echo "########## SMART status report summary for all drives ##########"
    echo ""
    echo "+------+---------------+----+-----+-----+-----+-------+-------+--------+------+----------+------+-------+"
    echo "|Device|Serial         |Temp|Power|Start|Spin |ReAlloc|Current|Offline |Seek  |Total     |High  |Command|"
    echo "|      |               |    |On   |Stop |Retry|Sectors|Pending|Uncorrec|Errors|Seeks     |Fly   |Timeout|"
    echo "|      |               |    |Hours|Count|Count|       |Sectors|Sectors |      |          |Writes|Count  |"
    echo "+------+---------------+----+-----+-----+-----+-------+-------+--------+------+----------+------+-------+"
) >> $logfile
for drive in $drives
do
    (
        $smartctl -A -i -v 7,hex48 /dev/$drive | \
        awk -v device=$drive -v tempWarn=$tempWarn -v tempCrit=$tempCrit -v sectorsCrit=$sectorsCrit \
        -v warnSymbol=$warnSymbol -v critSymbol=$critSymbol '\
       /Serial Number:/{serial=$3} \
       /Temperature_Celsius/{temp=$10} \
       /Power_On_Hours/{onHours=$10} \
       /Start_Stop_Count/{startStop=$10} \
       /Spin_Retry_Count/{spinRetry=$10} \
       /Reallocated_Sector/{reAlloc=$10} \
       /Current_Pending_Sector/{pending=$10} \
       /Offline_Uncorrectable/{offlineUnc=$10} \
       /Seek_Error_Rate/{seekErrors=("0x" substr($10,3,4));totalSeeks=("0x" substr($10,7))} \
       /High_Fly_Writes/{hiFlyWr=$10} \
       /Command_Timeout/{cmdTimeout=$10} \
       END {
           if (temp > tempCrit || reAlloc > sectorsCrit || pending > sectorsCrit || offlineUnc > sectorsCrit)
               device=device " " critSymbol;
           else if (temp > tempWarn || reAlloc > 0 || pending > 0 || offlineUnc > 0)
               device=device " " warnSymbol;
           seekErrors=sprintf("%d", seekErrors);
           totalSeeks=sprintf("%d", totalSeeks);
           if (totalSeeks == "0") {
               seekErrors="N/A";
               totalSeeks="N/A";
           }
           if (hiFlyWr == "") hiFlyWr="N/A";
           if (cmdTimeout == "") cmdTimeout="N/A";
           printf "|%-6s|%-15s| %s |%5s|%5s|%5s|%7s|%7s|%8s|%6s|%10s|%6s|%7s|\n",
           device, serial, temp, onHours, startStop, spinRetry, reAlloc, pending, offlineUnc, \
           seekErrors, totalSeeks, hiFlyWr, cmdTimeout;
       }'
    ) >> $logfile
done
(
    echo "+------+---------------+----+-----+-----+-----+-------+-------+--------+------+----------+------+-------+"
    echo ""
    echo ""
) >> $logfile
 
###### for each drive ######
for drive in $drives
do
    brand=`$smartctl -i /dev/$drive | grep "Model Family" | awk '{print $3, $4, $5}'`
    model=`$smartctl -i /dev/$drive | grep "Device Model" | cut -d':' -f2 | sed -e 's/^[ \t]*//g'`
    serial=`$smartctl -i /dev/$drive | grep "Serial Number" | awk '{print $3}'`
    (
        echo ""
        echo "########## SMART status report for $drive drive ($brand $model: $serial) ##########"
        $smartctl -n never -H -A -l error /dev/$drive
        $smartctl -n never -l selftest /dev/$drive | grep "# 1 \|Num" | cut -c6-
        echo ""
        echo ""
    ) >> $logfile
done
sed -i -e '/smartctl 6.3/d' $logfile
sed -i -e '/Copyright/d' $logfile
sed -i -e '/=== START OF READ/d' $logfile
sed -i -e '/SMART Attributes Data/d' $logfile
sed -i -e '/Vendor Specific SMART/d' $logfile
sed -i -e '/SMART Error Log Version/d' $logfile
echo "</pre>" >> $logfile
 
### Send report ###
cat $logfile | mail -h "Content-Type: text/html" -s "$subject" $email
rm $logfile