video card forwarding VE 6.4-4

It works stably only when there are no users on the virtual machines. It works when there are 5 users on each server. As soon as I let 15 users start such errors. Help me figure it out.
 
How was the RAM usage when and before it crashed? I'd guess the machine just ran out of RAM.
 
Before I configured ballooning, the memory of each VM was under 100 % in the proxmox web interface. In the virtual machines themselves in Windows, the memory load was normal. After configuring ballooning in the web interface, the memory load for each VM is displayed the same as in Windows. The total memory load of the node is 89%. Before the failure, the memory load of each VM does not exceed 50%. When a crash occurs, I did not notice that the memory load reached 100%. After setting up ballooning, the node behaves more stable, but until I let 15 people on the server. Before setting up ballooning, the node could hang without loading the VM.
 
Do you mean there is not enough memory for the virtual machines or the total for the node? In Windows of the VM, I didn't see the maximum memory load during the crash
 
Hi. Switched to the new version 6.4-4. The NOD lived a record time of 12 days. People from the servers were disconnected, but the Windows log of the VM's shows that the machines were still running. The node could be accessed via SSH. I wrote the reboot command in SSH and no reaction. In the WEB interface, all resources including the VM's were marked with a question mark. After I wrote the reboot command, there were lines about trying to stop the VM's. How to fix it?
 

Attachments

You must log in or register to reply here.
Top Bottom