Fix: DNS Resolution Failure

!Symptoms

• curl or wget fails with 'Could not resolve host'
• ping shows 'Name or service not known'
• Applications fail to connect to external services
• Browser shows DNS_PROBE_FINISHED_NXDOMAIN
• Kubernetes pods cannot resolve service names

?Root Causes

• /etc/resolv.conf has wrong or unreachable nameservers
• DNS server is down or overloaded
• Network configuration changed (DHCP lease renewal, VPN toggle)
• In Kubernetes: CoreDNS pods are down or misconfigured
• Firewall blocking UDP port 53 outbound
• systemd-resolved service is crashed

#Diagnosis Steps

1. 1Test DNS resolution: `nslookup google.com` or `dig google.com`
2. 2Check current DNS config: `cat /etc/resolv.conf`
3. 3Test with a known public DNS: `dig @8.8.8.8 google.com`
4. 4For Kubernetes: `kubectl get pods -n kube-system -l k8s-app=kube-dns`
5. 5Check if systemd-resolved is running: `systemctl status systemd-resolved`
6. 6Test connectivity to DNS port: `nc -zvu <dns-server> 53`

>Fix

1. 1Temporarily use public DNS: add `nameserver 8.8.8.8` to /etc/resolv.conf
2. 2Restart systemd-resolved: `systemctl restart systemd-resolved`
3. 3In Kubernetes, restart CoreDNS: `kubectl rollout restart deployment/coredns -n kube-system`
4. 4Fix firewall rules to allow outbound UDP/TCP 53
5. 5Flush DNS cache: `systemd-resolve --flush-caches` or `resolvectl flush-caches`

*Prevention

• Configure multiple DNS servers in resolv.conf for redundancy
• Monitor DNS resolution latency and failure rates
• In Kubernetes, ensure CoreDNS has adequate resources and replicas
• Use ndots and search domain settings appropriate for your environment
• Set up DNS health checks in your monitoring stack

Related Error Messages