Asm voting diskgroup created in normal redundancy size difference

[ETSystems (Systems Support ET)]

Hi,

I am trying to configure ASM checks for one of our RAC servers on CMK version: 2.1.0p19
But there is a size difference in the check created.
The allocated size for the voting disk group is 3*10GB but in check, it is showing only 15GB.
Any guess?

Thank you.

[ETSystems (Systems Support ET)]

Hi,

Can someone help me with this issue?
Have anyone faced this before?

[ETSystems (Systems Support ET)]

Still I’m waiting for valuable comments here. Any input would be much appreciated

[gstolz (Gerd Stolz)]

Can you share the output of the <<<oracle_asm_diskgroup>>> section in the check_mk_agent output?
(Please use code tags ;))

That might make troubleshooting easier

[ETSystems (Systems Support ET)]

hi @gstolz ,

Please find the details below,

<<<oracle_asm_diskgroup:sep(124)>>>
MOUNTED|NORMAL|VOTING|4096|1048576|10239|10239|9938|VOTING_0000|Y|REGULAR|0|8640000|1
MOUNTED|NORMAL|VOTING|4096|1048576|10239|10239|9925|VOTING_0001|Y|REGULAR|0|8640000|1
MOUNTED|NORMAL|VOTING|4096|1048576|10239|10239|9925|VOTING_0002|Y|REGULAR|0|8640000|1

Kindly note that these checks are created inside a cluster environment.

[gstolz (Gerd Stolz)]

hm… I guess the oracle_asm_diskgroup does some calculations with these but I’m not sure how that makes sense (then again; I know little to nothing about Oracle ASM)

@Rendanic can you explain what’s happening here and if it is correct?

Gerd

[Rendanic (Thorsten Bruhns)]

The checkresults is correct.
It is the way ASM works. The redundancy level is the critical prameter.

external redundancy => no data mirror
normal redundancy => each block has 1 mirror in a different failgroup
high redundancy => each block has 2 mirrors in different failgroups

It doesn’t matter how many failgroups are existing. You have to copy each block in a 2nd failgroup when normal redundancy is used.
Each failgroup should have the same size. Otherwise it is hard to calculate the useable space in the diskgroup.

The calculation for normal redundancy is:
Calculate the summary of all failgroups and devide it by 2.
=> 3*10GB / 2 = 15GB

That’s it for the space calculation.

The voting diskgroup is often is misunderstand in RAC environments…
We always need an unequal numer of failgroups for voting - the redundancy level defines the minimum number of failgroups for voting. (1 for external, 3 for normal and 5 for high)

Important! That does not affect the space calculation in the diskgroup itself!

Summary:
The check is fine. It is a misunderstand of the concept of data mirroring in ASM.