Memory usage of Linux VMs is a little misleading

  • 17 January 2018
  • 7 replies

Badge +1
Prism appears to report the memory usage of our Linux (Ubuntu) guests incorrectly.

The user interface shows that 94% of memory is used:

Whereas the linux agent is showing 21GB is available.

root@build1:/home/user# free -h total used free shared buff/cache availableMem: 31G 9.2G 292M 8.8M 21G 21GSwap: 11G 66M 11GTechnically, the memory is being used, but the os would allow the applications to use more RAM if they need to (see the below link for reasons why).

The "Available" memory information would be useful to display in order to allow administrators to easily see the actual usage of their VM requirements, and make judgements as to whether their VMs are over or underprovisioned.

FWIW our ubuntu version is:

root@build1:/user# lsb_release -aNo LSB modules are available.Distributor ID: UbuntuDescription: Ubuntu 16.04.3 LTSRelease: 16.04Codename: xenial

This topic has been closed for comments

7 replies

Userlevel 3
Badge +13
testworksau What is the AHV and AOS version you are using?
Badge +1
Userlevel 3
Badge +13
we are using the following command (from ahv) to find out the memory usage of the vm

virsh dommemstat
what does the output show for your vm?
Badge +1
I will have to get our Nutanix admins to run this - will get back to you!
Dear @Chandru , @testworksau

I have the some problem with Prism and my Linux VMs.
On my side, I have performed the command from the CVM :
virsh dommemstat

Please find hereafter my output :

virsh # dommemstat blablabla
actual 134217728
swap_in 97120
swap_out 1047372
major_fault 15077
minor_fault 1705330508
unused 4371388
available 131840524
rss 42328

Unfortunately, these values are not in accordance with the Prism information : which indicates 93% of Memory Usage.
So, judging from the virsh output, I expected ~ 98% of available memory....

So, it seems that Nutanix is taken into account the wrong information for the memory usage.

Any idea ?

Please don't hesitate to contact me directly at
Badge +1

Any updates on this one?
Linux is CENTOS 7

Nutanix - Version euphrates-5.5.7-stable LTS - Starter License

Nutanix UI is showing 99% memory usage for many of the Linux machines.

Running ‘top’ on the VM revels that 

top - 09:53:47 up 189 days, 1:39, 1 user, load average: 0.09, 0.16, 0.19
Tasks: 217 total, 1 running, 216 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.4 us, 0.4 sy, 0.0 ni, 99.2 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 57647192 total, 733548 free, 35421508 used, 21492136 buff/cache
KiB Swap: 10485756 total, 9193208 free, 1292548 used. 19010896 avail Mem

So Nutanix is considering the buff/cache as used?!

Is this a bug or a feature?


Badge +1

Wondering the same thing. Have CentOS 7 VM’s that Prism thinks are “Constrained”. Prism shows 92% and 93% memory used but it is all used as buff/cache.

# free

              total        used        free      shared  buff/cache   available

Mem:        1944592      194144      156036        8900     1594412     1567696

Swap:       1052668      154880      897788


I actually do not really see how Nutanix can detect this correct anyway unless you enable NGT and grant the CVM’s access to the VM’s network wise. I suspect without NGT memory usage statistics are useless.

Also asume “virsh dommemstat” is part of NGT.