Problem with ageneric-snmp-all-partitions (space discrepancy)

Icinga Director
1

Hello there.
We have a problem when we try to get information about available space using ageneric-snmp-all-partitions. We still have a Icinga 1 running which shows the correct available disk space, but Icinga 2 has on several different machines shown other numbers. The difference seems to be about 5%, whereas Icinga 2 shows that 5% more is available than what is actually available (and what Icinga 1 shows).

Here is one example from a NAS. This is what we can see in Icinga 2:
/: 52%
/dev/shm: 0%
/run: 2%
/sys/fs/cgroup: 0%
/tmp: 0%
/volume1: 86% (<90%) : OK

	Name			Used		Max			Warning		Critical
	/volume1		6,71 TiB	7,85 TiB	7,06 TiB	7,45 TiB
	/				1204,74 MiB	2,28 GiB	2,05 GiB	2,16 GiB
	/run			38,30 MiB	1888,00 MiB	1699,00 MiB	1794,00 MiB
	/tmp	   		1248,05 KiB	1888,00 MiB	1699,00 MiB	1794,00 MiB
	/dev/shm		120,01 KiB	1888,00 MiB	1699,00 MiB	1794,00 MiB
	/sys/fs/cgroup	0,00 B		1888,00 MiB	1699,00 MiB	1794,00 MiB

The command looks like this:

template Service "ageneric-snmp-all-partitions" {
    import "Active_Template_Check_Ausführung"

    check_command = "snmp-storage-linux"
    vars.snmp_crit = "95"
    vars.snmp_storage_name = "/"
    vars.snmp_warn = "90"
}

And these are the numbers that Icinga 1 shows us and which are indeed what is actually on the machine, it does use the “check_all_partitions” command though:

WARNING: /volume1:90%
/:56%
/run:7%
/tmp:5%
/dev/shm:5%
/sys/fs/cgroup:5%

This problem also does occur on more or maybe even all servers. I checked another Linux VM which is nearly empty, and while Icinga 1 shows around 88% on one partition and 5% on all others, Icinga 2 shows 84% and 0%-1% on all others, so again around 5% discrepancy.
Now I wasn’t the one who actually worked on Icinga 2, but I am the one who now has to fix it. I see that Icinga 2 also has a “check_disk” command. I am not completely sure why one command was used instead of another or of the “check_all_partitions” simply doesn’t exist anymore. Any ideas what may cause this discrepancy?

  • Director version (System - About): 1.8.0
  • Icinga Web 2 version and modules (System - About): 2.10.1
  • Icinga 2 version (icinga2 --version): r2.13.3-1)
  • Operating System and version: Debian 10
  • Webserver, PHP versions: Apache 2: 2.4.38-3+deb10u7 & php 7.3
2

Okay, I took another look at the data and while it’s still a bit iffy it seems that Icinga 2 is actually more correct than Icinga 1. The NAS in question really does have 86% used space (6,7 TB is all I can see) and not 90%.

However, on the Linux VM it’s a bit different. The VM itself shows 89%, Icinga 1 says it’s 88% and Icinga 2 says it’s only 84%. The 5% on the other partitions is however closer to 0% again.

So let me rephrase my question:
Any idea where this discrepancy comes from? I assume it may be an error in Icinga 1 now, but why would it show only 84% on the Linux VM when the VM itself shows 89% when looking at it with “df -h”?