The Symptom
You’re pinging localhost or a custom hostname you added to /etc/hosts and the request hangs or returns “Name or service not known”. The same name resolves fine on another machine, and dig or nslookup against a public DNS server works. The problem is isolated to the local resolver. In most modern distributions this points to a mis‑configured systemd‑resolved service or a stale cache.
Why systemd‑resolved Matters
systemd‑resolved is the default name‑resolution daemon on Fedora, Debian, Ubuntu, Arch, and many other distros since 2015. It replaces the older dnsmasq‑based /etc/resolv.conf approach and offers:
- Local hostname resolution via
/etc/hostsand/run/systemd/resolve/resolv.conf. - DNS over TLS (DoT) and DNSSEC validation.
- IPv6‑only or IPv4‑only mode.
- Link‑local and multicast DNS (mDNS) support.
Because it runs as a system service, a mis‑state in its configuration or a stale cache can silently break local name resolution, even though /etc/hosts looks fine.
1. Verify the Basic State
# Check the current /etc/hosts entry
grep -E '^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+' /etc/hosts
# Verify systemd-resolved is active
systemctl status systemd-resolved.service
If the service is inactive or failed, start it:
sudo systemctl enable --now systemd-resolved
Common Pitfall: /etc/hosts Not Read
systemd‑resolved reads /etc/hosts only if the file is owned by root and has 0644 permissions. A 0600 or 0666 file will be ignored.
sudo chown root:root /etc/hosts
sudo chmod 0644 /etc/hosts
2. Inspect the Resolver Configuration
systemd‑resolved exposes its state via systemd-resolve --status. Look for:
- Current DNS servers – are they reachable?
- DNSSEC – is it enabled? A mis‑configured DNSSEC policy can cause failures for names that do not have valid signatures.
- Link‑specific settings – sometimes a network interface is set to use a different DNS server.
systemd-resolve --status
Example Output
Global
DNS Servers: 192.168.1.1
DNSSEC NTA: 0.0.0.0/0
DNSSEC supported: yes
DNSSEC enabled: yes
DNSSEC trust anchors: 2
DNSSEC key type: RSA/SHA-256
DNSSEC key size: 2048
If the DNS servers are unreachable, the resolver will fall back to the local cache, which may not contain the host you added.
3. Flush the Cache
Sometimes the cache holds an old negative answer. Flush it and try again.
sudo systemd-resolve --flush-caches
After flushing, test:
ping -c 1 myhost.local
If the ping still fails, move to the next step.
4. Check the Link‑Local and mDNS Settings
systemd‑resolved can be configured to ignore /etc/hosts if the --no-hosts flag is set in the unit file or via the systemd-resolved.conf file.
sudo systemctl edit systemd-resolved
If you see:
[Resolve]
NoHosts=yes
remove or comment it out, then reload:
sudo systemctl daemon-reload
sudo systemctl restart systemd-resolved
5. Verify /etc/resolv.conf Symlink
systemd‑resolved usually creates a symlink:
ls -l /etc/resolv.conf
It should point to /run/systemd/resolve/stub-resolv.conf. If it points elsewhere (e.g., /etc/resolve.conf or a static file), the resolver may not be used.
sudo ln -sf /run/systemd/resolve/stub-resolv.conf /etc/resolv.conf
6. Test with dig and nslookup
dig bypasses the local resolver and queries the DNS servers directly. This helps confirm whether the issue is local or remote.
dig @192.168.1.1 myhost.local
If this returns NXDOMAIN, the DNS server is the problem. If it returns a valid A record, the local resolver is still mis‑behaving.
7. Disable systemd‑resolved Temporarily
If you need a quick workaround, you can disable systemd‑resolved and fall back to the classic /etc/resolv.conf approach.
sudo systemctl disable --now systemd-resolved
sudo rm /etc/resolv.conf
sudo touch /etc/resolv.conf
echo "nameserver 8.8.8.8" | sudo tee /etc/resolv.conf
This is not a permanent fix, but it lets you keep working while you investigate.
8. Common Security Implications
| Issue | Risk | Mitigation |
|---|---|---|
systemd‑resolved disabled | DNS spoofing, MITM | Keep systemd‑resolved enabled and use DNSSEC. |
NoHosts=yes | Local hosts bypassed, potential for host spoofing | Ensure NoHosts=no unless you have a specific reason. |
| Stale cache | False positives in security monitoring | Flush cache regularly or enable CacheSize=0 for no caching (not recommended for performance). |
9. Advanced: Using systemd-resolve Options
You can override DNS settings per‑link without editing /etc/netplan or NetworkManager:
sudo systemd-resolve --interface eth0 --set-dns 10.0.0.1
sudo systemd-resolve --interface eth0 --set-domain example.com
These commands are useful when you have multiple networks with different DNS policies.
10. Persistent Configuration
Edit /etc/systemd/resolved.conf to make permanent changes:
[Resolve]
DNS=192.168.1.1 8.8.8.8
FallbackDNS=1.1.1.1
Domains=~.
LLMNR=no
MulticastDNS=no
DNSSEC=yes
After editing:
sudo systemctl restart systemd-resolved
11. Troubleshooting Checklist
| Step | Command | What to look for |
|---|---|---|
| 1 | systemctl status systemd-resolved | Active, no errors |
| 2 | systemd-resolve --status | DNS servers reachable, DNSSEC enabled |
| 3 | cat /etc/hosts | Correct syntax, permissions |
| 4 | systemd-resolve --flush-caches | Cache cleared |
| 5 | dig @<dns> <name> | Direct DNS response |
| 6 | ping <name> | Local resolution success |
If all checks pass but resolution still fails, consider:
- A corrupted
systemd-resolvedbinary (rare, but possible after a kernel upgrade). - A network interface mis‑configured to use a different DNS server.
- SELinux or AppArmor blocking access to
/etc/hosts(check audit logs).
12. Real‑World Example
Scenario
A Debian 12 server in a homelab had a custom hostname media.local added to /etc/hosts. After a kernel upgrade, ping media.local hung. The admin ran:
systemctl status systemd-resolved
The service was active, but systemd-resolve --status showed DNSSEC enabled: no. The admin had previously set DNSSEC=yes in /etc/systemd/resolved.conf. The file had been overwritten by the upgrade.
Fix
sudo nano /etc/systemd/resolved.conf
# Add:
[Resolve]
DNSSEC=yes
Then:
sudo systemctl restart systemd-resolved
sudo systemd-resolve --flush-caches
ping media.local
The ping succeeded. The admin added a comment in the file to remember why DNSSEC was re‑enabled.
13. When to Bypass systemd‑resolved
In environments where you need deterministic DNS resolution (e.g., embedded devices, minimal containers), you might prefer a static /etc/resolv.conf. In such cases:
- Disable
systemd‑resolved. - Create a static
/etc/resolv.confwith the required nameservers. - Ensure no other service rewrites the file (e.g.,
NetworkManager).
This approach sacrifices DNSSEC and DoT but can simplify troubleshooting in constrained environments.
14. Keeping the System Secure
- Enable DNSSEC:
systemd-resolvedvalidates signatures, reducing DNS spoofing risk. - Use DoT: Configure
systemd-resolvedto use a DoT provider (e.g., Cloudflare 1.1.1.1:853). - Restrict mDNS: Disable
MulticastDNS=yesunless you need local service discovery. - Audit logs: Monitor
/var/log/systemd/journalforsystemd-resolveerrors.
15. Quick Reference Commands
# Check status
systemctl status systemd-resolved
# View config
systemd-resolve --status
# Flush cache
sudo systemd-resolve --flush-caches
# Edit config
sudo nano /etc/systemd/resolved.conf
# Restart service
sudo systemctl restart systemd-resolved
# Verify /etc/hosts permissions
stat -c "%U %G %a" /etc/hosts
See also
- Application-Level Firewalls
- Fixing a Crashed Btrfs Subvolume After Unexpected Reboot
- Recovering a Kernel Panic After Adding a New Initramfs Module
- Multi‑jump bastion access without storing keys: practical ssh‑agent forwarding
- Rolling Back a Broken Flatpak Update
- Fixing systemd-resolved Ignoring /etc/hosts