ESXi Using Too Much RAM with No VMs Running

[vikest]

Hey Team, a fresh ESXi install with 5 VMs including vCenter is using up to 30GBs of RAM with all of the VMs shut down. Is there a reason for this or is this something that needs addressing? Reinstalling the whole thing did not fix the problem. This is on the latest 8.0U1a version of both ESXi and vCenter.

Thanks in advance!

[AnaghB]

Hello @vikest ,

Please check the total usage of the 5 VMs that are deployed and post that check the usage of the Esxi. See what is the difference and let me know.

[vikest]

Thanks Anagh, ESXi uses almost 30GB without any VMs running. When the VMs are started, they use 25GB on top of those 30GB

[AnaghB]

Hello @vikest ,

Please check if those 5 VMs that are deployed by default are using any RAM. If yes that how much?

[vikest]

When the VMs are powered off, they're not utilizing any RAM of course, but yet the ESXi host keeps using 30GB

[Sachchidanand]

Login to the ESXi CLI mode using ssh and run the "esxtop" and then press "m" on your keyboard to switch to memory pane and see which process is consuming the most memory. 

Regards,

Sachchidanand

[vikest]

Thank you that was very helpful! Is it being reserved? When I fire up 5 VMs, it uses pretty close to all of it. As a test, I cloned them and fired up 55 VMs, I didn't notice much a difference in utilization, and the machines seemed ok. Reinstalling ESXi from the ground up gives me the same results after installing vCenter. But here, vCenter is powered off.

[Sachchidanand]

Please go through the following document which defines all the terms in the output of "esxtop" command and having some Q&A to help understanding it. Please see the memory section.

https://communities.vmware.com/t5/Storage-Performance/Interpreting-esxtop-Statistics/ta-p/2776936

Regards,

Sachchidanand 

[Tibmeister]

What does the rest of the environment look like?  vSAN, NFS, iSCSI?  NSX?  I've seen a standalone ESXi host fresh installed take up to 15% of available RAM to run just basic services.  In your case, you're looking at close to 50%, so there's something else in the environment that is requiring this amount of RAM.