Description of problem: I've configured dnsmasq to resolve my local libvirt VMs through using a NetworkManager file /etc/NetworkManager/dnsmasq.d/00-libvirt.conf: -------- 8< ---------------- 8< ---------------- 8< ---------------- 8< -------- server=/libvirt/192.168.122.1 -------- 8< ---------------- 8< ---------------- 8< ---------------- 8< -------- This always worked well until today where I realized dnsmasq was updated yesterday from 2.86-2.fc35 to 2.86-5.fc35. Since then, after connecting to Red Hat VPN, the name resolution stops functioning: 1. Before connecting to VPN -------- 8< ---------------- 8< ---------------- 8< ---------------- 8< -------- $ ping -c 1 vm-rhel8 PING vm-rhel8.libvirt (192.168.122.86) 56(84) bytes of data. -------- 8< ---------------- 8< ---------------- 8< ---------------- 8< -------- 2. After connecting to VPN -------- 8< ---------------- 8< ---------------- 8< ---------------- 8< -------- $ ping -c 1 vm-rhel8 ping: vm-rhel8: No address associated with hostname -------- 8< ---------------- 8< ---------------- 8< ---------------- 8< -------- Stracing main dnsmasq instance, it looks like the request is not forwarded to dnsmasq instance dealing with my "libvirt" domain. Restarting NetworkManager (which kills the VPN connection) makes it work again, until I connect to the VPN. Downgrading dnsmasq to 2.86-2.fc35 permanently solves the issue: it continues resolving after connecting to the VPN. Version-Release number of selected component (if applicable): 2.86-5.fc35 How reproducible: Always Steps to Reproduce: see above Actual results: No resolution of my "libvirt" domain Expected results: Resolution
Created attachment 1866089 [details] strace when VPN is not active (resolution works fine)
Created attachment 1866090 [details] strace just after VPN becomes active (no resolution anymore)
It seems this is duplicate to bug #2061944. It should be fixed soon. *** This bug has been marked as a duplicate of bug 2061944 ***