Check Disk Unix I/O

[murawweh.daher]

Hello,

i installed nrpe agent on solaris10 and want to check disk i/o, i tried to use check_iostat plugin but everytime have error as below.
what is the best way to check disk load for solaris 10?

./check_iostat -h
./check_iostat: function: not found
-e
This plugin shows the I/O usage of the specified disk, using the iostat external program.
It prints three statistics: Transactions per second (tps), Kilobytes per second
read from the disk (KB_read/s) and and written to the disk (KB_written/s)

./check_iostat:
-d <disk> Device to be checked (without the full path, eg. sda)
-c <tps>,<read>,<wrtn> Sets the CRITICAL level for tps, KB_read/s and KB_written/s, respectively
-w <tps>,<read>,<wrtn> Sets the WARNING level for tps, KB_read/s and KB_written/s, respectively

./check_iostat: -1: bad number

[rkennedy]

What version of the check_io script are you running?

[murawweh.daher]

!/bin/sh
#
# Version 0.0.2 - Jan/2009
# Changes: added device verification
#
# by Thiago Varela - [email protected]

please find attached file (please remove .txt just for uploading)

[tgriep]

If you edit that plugin and remove the following lines from it, that message will go away.

Code: Select all

function help {
echo -e "\n\tThis plugin shows the I/O usage of the specified disk, using the iostat external program.\n\tIt prints three statistics: Transactions per second (tps), Kilobytes per second\n\tread from the disk (KB_read/s) and and written to the disk (KB_written/s)\n\n$0:\n\t-d <disk>\t\tDevice to be checked (without the full path, eg. sda)\n\t-c <tps>,<read>,<wrtn>\tSets the CRITICAL level for tps, KB_read/s and KB_written/s, respectively\n\t-w <tps>,<read>,<wrtn>\tSets the WARNING level for tps, KB_read/s and KB_written/s, respectively\n"
	exit -1
}

Solaris's command shell doesn't work with functions used in scripts.

Try that and see if it works for you.

[murawweh.daher]

./check_iostat -d / -w 1,2,3 -c 2,3,4
ERROR: Device incorrectly specified
./check_iostat: help: not found
./check_iostat: test: unknown operator ==
./check_iostat: [[: not found
./check_iostat: [[: not found
./check_iostat: [[: not found
./check_iostat: : cannot execute
./check_iostat: test: unknown operator 2 | bc
./check_iostat: : cannot execute
./check_iostat: test: unknown operator 1 | bc
OK - I/O stats tps= KB_read/s= KB_written/s= | 'tps'=; 'KB_read/s'=; 'KB_written/s'=;

[rkennedy]

./check_iostat -d / -w 1,2,3 -c 2,3,4
ERROR: Device incorrectly specified
./check_iostat: help: not found
./check_iostat: test: unknown operator ==
./check_iostat: [[: not found
./check_iostat: [[: not found
./check_iostat: [[: not found
./check_iostat: : cannot execute
./check_iostat: test: unknown operator 2 | bc
./check_iostat: : cannot execute
./check_iostat: test: unknown operator 1 | bc
OK - I/O stats tps= KB_read/s= KB_written/s= | 'tps'=; 'KB_read/s'=; 'KB_written/s'=;

I believe it's looking for a mount, not just /.

Code: Select all

-d <disk> Device to be checked (without the full path, eg. sda)

[murawweh.daher]

i tried without full path and same issue.

[lmiltchev]

Use "sda", instead of "/". Example:

Code: Select all

[root@192 tmp]# df -h
Filesystem      Size  Used Avail Use% Mounted on
rootfs          9.4G  6.1G  3.3G  66% /
devtmpfs        2.0G  140K  2.0G   1% /dev
tmpfs           2.0G     0  2.0G   0% /dev/shm
/dev/sda1       9.4G  6.1G  3.3G  66% /

[root@192 tmp]# ./check_iostat.sh -d sda1 -w 1,2,3 -c 2,3,4
CRITICAL - I/O stats tps=7.46 KB_read/s=3.35 KB_written/s=174.58 | 'tps'=7.46; 'KB_read/s'=3.35; 'KB_written/s'=174.58;

[murawweh.daher]

i test on Linux and worked fine.
./check_iostat -c 20,20,20 -w 10,10,10 -d sda
CRITICAL - I/O stats tps=232.33 KB_read/s=238.10 KB_written/s=5615.89 | 'tps'=232.33; 'KB_read/s'=238.10; 'KB_written/s'=5615.89;


but on Solaris 10 didn't work.

[rkennedy]

Can you post the full input / output when you attempt it on the Solaris machine?