Proxmox showing high usage memory on Dashboard for Truenas(VM)

[sukarechhe]

Dear Proxmox Community,

I am reaching out to seek assistance regarding a recurring issue with my Proxmox physical server (Proxmox Virtual Environment 8.2.2). The Proxmox dashboard displays memory usage information for my TrueNAS VM (TrueNAS-13.0-U6.2) that does not align with the actual usage reported within the TrueNAS system.

Details of the Problem:​

1. The Proxmox dashboard consistently shows significantly higher memory usage for the TrueNAS VM than the usage reported inside the VM itself.
2. When the memory usage on the Proxmox dashboard reaches its maximum, the TrueNAS VM begins to misbehave:
   • SMB services stop functioning.
   • The Proxmox host dashboard becomes inaccessible on port 8006.

I have ensured that the memory allocation for the TrueNAS VM is configured correctly. Additionally, I have installed the qemu-guest-agent service within the VM to improve reporting accuracy. However, the discrepancy in memory usage persists, leading to crashes when the memory peaks.

I would greatly appreciate guidance on the following:

1. Understanding why there is a mismatch in memory usage reporting between Proxmox and TrueNAS, even with the qemu-guest-agent service installed.
2. Steps to prevent the crashes caused by memory-related issues.

For reference, I am attaching screenshots to provide more detailed insights into the problem.

Thank you for your time and support. I look forward to any suggestions or recommendations you can provide.

Best regards,

 

[cwt]

Real RAM usage is often „blurred“ by caching which is pretty normal for nearly every virtualized OS. Proxmox itself also uses RAM for caching, especially when ZFS is used.

How did you setup your TrueNAS vHDDs? Passthrough or local storage? What‘s not recommended is the usage of ZFS inside VMs on top of an existing ZFS dataset in Proxmox.