I had the same issue: small packets were lost on vmware NAT interface (vmnet8). DNS resolution was slow, because (linux) client had to re-send requests after timeout. Sometimes a packet was not transfered even after 3 retries.
The problem was fixed by restarting the "VMware NAT Service".
Host OS: windows 10
VMware Workstation 15 Pro 15.5.1 build-15018445
Guest OS: ubuntu 18.04.3.
Restating the service:
1. Press windows or start button.
2. Type "services.msc" and press "ctrl+shift+enter" to run as admin.
3. Select "VMware NAT Service" and choose Action->Restart from dropdown menu.
The same fix effect can be triggered by changing the NAT network DNS settings:
- VMware workstation main menu -> Edit -> Virtual network editor
- Virtual network editor window will open. Enable administrator mode by pressing "Change setings" in bottom right corner of the modal window.
- Select "VMnet8 (NAT)" and press "NAT settings".
- Press "DNS Settings"
- Change "Policy" to burst and reduce timeout to 1 second. Optionally increase retries.
- The "VMware NAT Service" will be automatically restarted after pressing OK in the Virtual network editor window.
This should fix DNS until the windows network adapters would change. For example, "VMware NAT Service" may go crazy again after reconnecting from one wifi to another.
Guest settigns that were checked and did NOT help to fix the issue:
- switch from "PCnet/PCI II" to "Intel PRO/1000" network adapter by adding this to .vmx: ethernet0.virtualDev = "e1000"
- manually select another DNS server: sed -iE 's/^nameserver .*$/nameserver 1.1.1.1/' /etc/resolv.conf
- disable all network card offloading: ethtool -K ens33 rx off tx off sg off tso off gso off rxvlan off gro off; ethtool --show-offload ens33 | grep ': on'
