VM to VM lag

[Zack Coffey]

We have a couple BIND9 servers on Ubuntu server setup on different Proxmox hosts.

Doing a very simple test of 'dig' from a separate host gives a 0-1ms response time. However, doing the same test from a new VM on the same Proxmox host as one of the BIND9 servers takes 22-37ms.

It's on the same physical machine, why would it take any longer than a separate physical machine?

 

[wolfgang]

Hi,

it is hard to say with so few information but I would check the network.

What are the result of ping?

 

[wbumiller]

What's the exact configuration of the two involved VMs and what does the network connecting them look like? (topology, configuration, firewall, traffic limits...)

 

[Zack Coffey]

Well, problem seems to be gone now. For the record, it's a mix of Ubuntu 14.04 and 16.04. Very simple setup, no firewalls, no traffic limits. The one VM I was seeing the problem with, both VM's were sitting on the same Proxmox host with, what I would consider, a simple setup. Each VM has 1 network connection via bridge to the rest of the network, each has a static IP in the same subnet. The other machine that's physical is on the same switch as the Proxmox host, but it's a physical machine running Ubuntu instead of a VM.

Again, it seems to be fine now. Pings and digs are showing 1ms max, average is sub 1ms response. The 1 VM guest I was seeing the problem on, I ran updates and rebooted it, maybe that made some magic happen.