LinuxQuestions.org - script to send mail when raid fails

- Linux - Newbie (https://www.linuxquestions.org/questions/linux-newbie-8/)

- - script to send mail when raid fails (https://www.linuxquestions.org/questions/linux-newbie-8/script-to-send-mail-when-raid-fails-929494/)

script to send mail when raid fails

Hi..All,
I have configured RAID1(software raid). And since I am newbie in shell scripting, I need shell script to send mail when any of the raid array get fails. It would be very appreciated if anybody already having the script and post it here. or please help me, how can I cut "[" "]" from the below command's output. The raid status is stored in mdstat file.

Code:

[ananth@localhost ~]$ cat mdstat | grep "blocks" | awk '{print $4'}

[UU]

[UU]

[UU]

[UU]

[UU]

[UU]

I'll expect the output as only UU instead of [UU].

First:
You don't need to cat AND grep - grep can read by itself - you just do:

Code:

grep "blocks" /proc/mdstat

Second awk by default uses white space (spaces/tabs etc...) for field delimiter so $4 would be the fourth field. Since there is no white space betwee "[" "UU" and "]" it is all considered one field.

You could try solving it by changing the delimiter twice assuming that "[" and "]" appear only once in the output:

Code:

grep "blocks" /proc/mdstat |awk -F[ '{print $2}' |awk -F] '{print $1}'

Notice your field is now $2 in the first awk because it is no longer delimited by white space and the "UU]" is the second thing after the "[" delimiter. Similarly in the second awk the field is $1 because "UU" is the first thing before the "]" delimiter.

You could perhaps simplify the above by using sed the command rather than dual awks.

Quote:

sed 's/^.*$UU$.*$/\1/'

This one liner works.

Thanks MensaWater
Sorry, I think I confused you by grep and print the output from only mdstat file instead /proc/mdstat file.

Code:

$grep "blocks" /proc/mdstat |awk -F[ '{print $2}' |awk -F] '{print $1}'

2/2

2/2

2/2

2/2

2/2

2/2

Since It was not giving output as "UU", I edited as below

Code:

$grep "blocks" /proc/mdstat |awk -F[ '{print $3}' |awk -F] '{print $1}'

UU

UU

UU

UU

UU

UU

Super, This is what I had expected from you. Here is the script, still giving the error but saying RAID OKAY.

Code:

$cat RAID

#!/bin/bash

status=( $(grep "blocks" /proc/mdstat |awk -F[ '{print $3}' |awk -F] '{print $1}' ) )

if [ "${!status[@]}" -ne "UU" ]; then

echo RAID FAILED!!!

else

echo RAID OKAY.

fi

Code:

$bash RAID

RAID: line 3: [: too many arguments

RAID OKAY.

Please help me that how can I get rid this error.

Thanks cbtshare,

Code:

$grep "blocks" /proc/mdstat | awk '{print $4'} | sed 's/^.*\(UU\).*$/\1/'

UU

UU

UU

UU

UU

UU

Super working fine.

Quote:

Originally Posted by niharikaananth (Post 4603594)

Thanks cbtshare,

Code:

$grep "blocks" /proc/mdstat | awk '{print $4'} | sed 's/^.*\(UU\).*$/\1/'

UU

UU

UU

UU

UU

UU

Super working fine.

Hi..Masters,
Here is a small script, still I am unable to ignore the errors.

Code:

cat check_raid_array

#!/bin/bash

status=( $(grep "blocks" /proc/mdstat | awk '{print $4'} | sed 's/^.*\(UU\).*$/\1/' ) )

if [ ${status[i]} -ne "UU" ]; then

echo RAID FAILED 

else

echo RAID OKAY

fi

If I run the script I am getting the error as below.

Code:

bash check_raid_array 

check_raid_array: line 3: [: UU: integer expression expected

RAID OKAY

Please help me to correct my mistakes.

-ne is a numeric comparison operator. You want the string comparison operator !=

Quote:

Originally Posted by catkin (Post 4604200)

-ne is a numeric comparison operator. You want the string comparison operator !=

Thanks catkin,

Code:

#!/bin/bash

status=( $(grep "blocks" /proc/mdstat | awk '{print $4'} | sed 's/^.*\(UU\).*$/\1/' ) )

if [ "${status[i]}" != "UU" ]; then

echo RAID FAILED 

else

echo RAID OKAY

fi

Code:

[root@localhost ananth]#bash check_raid_array 

RAID OKAY

[root@localhost ananth]#echo $?

0

Working fine, But it is not showing as RAID FAILED while $status is U_ or UF or only U in any of the raid array. For example cat /proc/mdstat output is saved in a file and changed the md4 status from "UU" to "U_", Please see the below example

Code:

[root@localhost ananth]# cat mdstat 

Personalities : [raid1] 

md0 : active raid1 sdb1[1] sda1[0]

      513984 blocks [2/2] [UU]

      

md2 : active raid1 sdb3[1] sda3[0]

      2096384 blocks [2/2] [UU]

      

md3 : active raid1 sdb5[1] sda5[0]

      1469824 blocks [2/2] [UU]

      

md4 : active raid1 sdb6[1] sda6[0]

      497856 blocks [2/2] [U_]

      

md5 : active raid1 sdd1[1] sdc1[0]

      20964672 blocks [2/2] [UU]

      

md1 : active raid1 sdb2[1] sda2[0]

      15358016 blocks [2/2] [UU]

And also modified the script to grep "blocks" from mdstat file instead of /proc/mdstat.

Code:

#!/bin/bash

status=( $(grep "blocks" mdstat | awk '{print $4'} | sed 's/^.*\(UU\).*$/\1/' ) )

if [ "${status[i]}" != "UU" ]; then

echo RAID FAILED 

else

echo RAID OKAY

fi

But if I run the script, still it is showing as "RAID OKAY" instead of RAID FAILED, which means I think it is reading only md0 and not reading other md devices.

Code:

[root@localhost ananth]# bash check_raid_array 

RAID OKAY

Please help me that what can be done, to read every raid array result.

because the

Quote:

if [ "${status[i]}" != "UU" ]; then

line makes no-sense really.I guess you wanna put the contents in an array and the the cycle through it ,but i doesn't increase.you'd have to let "i =+ 1" , but I'd use a while loop with the nested If.

Thanks cbtshare,
Since I am newbie in shell scripting I am unable to modify the script with "let". So could you please post the modified script with while loop.

Assuming the grep awk line does what I think ie return status values like UU, U_ etc, how about something like

Code:

#!/bin/bash

for status in $(grep "blocks" mdstat | awk '{print $4'} | sed 's/^.*\(UU\).*$/\1/' ) 

do 

    if [[ "$status" != "UU" ]]

    then

        echo RAID FAILED 

    else

        echo RAID OKAY

    fi

done

For more sophistication, you'd also pull out the mdX values, so you know which one has failed;
not strictly reqd, but nice to have.

BTW, here are some good bash links
http://rute.2038bug.com/index.html.gz
http://tldp.org/LDP/Bash-Beginners-G...tml/index.html
http://www.tldp.org/LDP/abs/html/

chrism01's code does it , try that :)

Thank you very much chrism01,
Sorry for the late response. Could you please help me to print failed devices using awk/sed? For example

Code:

cat mdstat

RAID status

Personalities : [raid1] 

md0 : active raid1 sdb1[1] sda1[0]

      1020032 blocks [2/2] [__]

      

md2 : active raid1 sdb3[1] sda3[0]

      20482752 blocks [2/2] [UU]

      

md3 : active raid1 sdb5[1] sda5[0]

      537117056 blocks [2/2] [UF]

      

md4 : active raid1 sdb6[1] sda6[0]

      315741376 blocks [2/2] [UU]

      

md5 : active raid1 sdd1[1] sdc1[0]

      976759936 blocks [2/2] [_U]

      

md1 : active raid1 sdb2[1] sda2[0]

      102398208 blocks [2/2] [U_]

      

unused devices: <none>

Here I want to print out only failed devices which are in red color. So it would be appreciated if you post the awk or sed commands to print only failed device and by excluding those raid array which are good i.e UU.