check_snmp_mem vs check_xi_service_snmp_linux_storage

[mbeebe]

Hello,

Can someone please explain the difference between check_xi_service_snmp_linux_storage and check_snmp_mem as far as memory monitoring goes?

The reason I ask is that since upgrading to 5.6.7, check_snmp_mem is retuning negative values for memory used:

Nagios shows: Ram : -49%, Swap : 5% : > 90, 5 : WARNING

Actual box shows (with free -m):
total used free shared buff/cache available
Mem: 7821 1719 163 2376 5938 3427
Swap: 4095 216 3879

However, our check_xi_service_snmp_linux_storage also reports incorrectly:
(this is a check from a different box)

Nagios: Physical memory: 95%used(3611MB/3792MB) (>95%) : WARNING

Actual box shows (with free -m):
total used free shared buff/cache available
Mem: 3792 210 175 184 3406 3117

Which snmp-based memory check should we be using and what's the difference between the two?

Thanks,

-- Mike Beebe

[scottwilkerson]

check_snmp_mem was not updated in any recent release (not for many years), so I'm not sure why that would be returning negative numbers, what is the full command you are using for that?

However, our check_xi_service_snmp_linux_storage also reports incorrectly:
(this is a check from a different box)

Nagios: Physical memory: 95%used(3611MB/3792MB) (>95%) : WARNING

Actual box shows (with free -m):
total used free shared buff/cache available
Mem: 3792 210 175 184 3406 3117

this one looks pretty close based on the calculation it uses

Code: Select all

Total - Free = Used
3792 - 175 = 3617

Plugin reported 3611MB

[mbeebe]

check_snmp_mem was not updated in any recent release (not for many years), so I'm not sure why that would be returning negative numbers, what is the full command you are using for that?

However, our check_xi_service_snmp_linux_storage also reports incorrectly:
(this is a check from a different box)

Nagios: Physical memory: 95%used(3611MB/3792MB) (>95%) : WARNING

Actual box shows (with free -m):
total used free shared buff/cache available
Mem: 3792 210 175 184 3406 3117

this one looks pretty close based on the calculation it uses

Code: Select all

Total - Free = Used
3792 - 175 = 3617

Plugin reported 3611MB

This issue for me is free vs. available. There's ~3gigs of memory available, even 'tho "free" memory reports as 175megs. The check should be looking at available memory, not free.

Is there a way to modify the alert so that it reads available memory and not free?

-- Mike Beebe

[scottwilkerson]

I don't know of a SNMP plugin that does this, but I do know it is possible with the XI Linux Agent used in the Linux Wizard and NCPA Agent using the NCPA Wizard

[mbeebe]

check_snmp_mem was not updated in any recent release (not for many years), so I'm not sure why that would be returning negative numbers, what is the full command you are using for that?

check_snmp_mem!90,5!95,10!!!!!!

[mbeebe]

I don't know of a SNMP plugin that does this, but I do know it is possible with the XI Linux Agent used in the Linux Wizard and NCPA Agent using the NCPA Wizard

Scott,

Would the same be true of the disk size check agent? Right now, it doesn't correctly return free space, instead it sends the difference between total space and space used which can come out as many gigs, even when there's actually no space available. This bit us this morning.

-- Mike Beebe

[scottwilkerson]

Scott,

Would the same be true of the disk size check agent? Right now, it doesn't correctly return free space, instead it sends the difference between total space and space used which can come out as many gigs, even when there's actually no space available. This bit us this morning.

-- Mike Beebe

I'm not familiar with this happening with disk space checks, can you give an example command you are using and the output?

[mbeebe]

Scott,

Would the same be true of the disk size check agent? Right now, it doesn't correctly return free space, instead it sends the difference between total space and space used which can come out as many gigs, even when there's actually no space available. This bit us this morning.

-- Mike Beebe

I'm not familiar with this happening with disk space checks, can you give an example command you are using and the output?

check_xi_service_snmp_linux_storage! -C yeahright --v2c -m "^/filesharename$" -w 90 -c 95 -f

Result:
/filesharename: 46%used(46398MB/100661MB)

Actual (df -hl):
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/vg_filesystem 99G 46G 48G 49% /filesharename

As you can see, the check reports 46% used vs. df reporting 49% used.

-- Mike Beebe

[scottwilkerson]

this is bizarre, I've never seen this before, the command for that check just runs

Code: Select all

/usr/local/nagios/libexec/check_snmp_storage_wizard.pl -H xxx.xxx.xxx.xxx -C yeahright --v2c -m "^/filesharename$" -w 90 -c 95 -f

This plugin pulls the data directly from what the system reports via SNMP

One thing of note on your df -hl command

Filesystem Size Used Avail Use% Mounted on
/dev/mapper/vg_filesystem 99G 46G 48G 49% /filesharename

Code: Select all

46G + 48G = 94G NOT 99G

What is up with that? Why is your df -hl command not adding up?

[mbeebe]

this is bizarre, I've never seen this before, the command for that check just runs

Code: Select all

/usr/local/nagios/libexec/check_snmp_storage_wizard.pl -H xxx.xxx.xxx.xxx -C yeahright --v2c -m "^/filesharename$" -w 90 -c 95 -f

This plugin pulls the data directly from what the system reports via SNMP

One thing of note on your df -hl command

Filesystem Size Used Avail Use% Mounted on
/dev/mapper/vg_filesystem 99G 46G 48G 49% /filesharename

Code: Select all

46G + 48G = 94G NOT 99G

What is up with that? Why is your df -hl command not adding up?

It's a Linux'ism: 5% of space is reserved for the root user so it can write to a volume even if it's "full".