Time is a hidden dependency of almost everything on a server: TLS certificates fail, log analysis becomes guesswork, database replication drifts and Kerberos refuses to authenticate when clocks disagree. RHEL 9 solves this with Chrony, and learning to configure Chrony on RHEL 9 takes only a few minutes. This guide shows how to configure Chrony as an NTP client, how to run your own internal time server for a LAN, and how to verify synchronization with chronyc.
Prerequisites – How to Configure Chrony on RHEL
- A server running RHEL 9, CentOS Stream 9, AlmaLinux 9 or Rocky Linux 9
- Root or sudo access
- For the server part: a static IP address and UDP port 123 reachable by your clients
Step 1: Install Chrony on RHEL 9 – How to Configure Chrony on RHEL
Chrony is installed by default on most RHEL 9 systems. If it is missing, install it and make sure the service is enabled:
dnf install -y chrony
systemctl enable --now chronyd
systemctl status chronyd
The daemon is called chronyd and the command-line control tool is chronyc. Note that Chrony replaces the older ntpd; the two should never run on the same machine.
Step 2: Configure Chrony on RHEL 9 as a Client
The configuration lives in /etc/chrony.conf. The only lines most clients need to touch are the time sources. Replace the default pool with servers that make sense for your region or your organisation:
# Use public NTP servers from the pool project
pool vn.pool.ntp.org iburst
pool time.google.com iburst
# Record the rate at which the clock gains/loses time
driftfile /var/lib/chrony/drift
# Step the clock if the adjustment is large in the first updates
makestep 1.0 3
# Enable kernel synchronisation of the real-time clock
rtcsync
The iburst option sends a quick burst of packets at startup so the clock converges in seconds instead of minutes. After editing, restart and check:
systemctl restart chronyd
chronyc tracking
chronyc sources -v
chronyc tracking shows your current offset under System time — on a healthy client it should be a tiny fraction of a second. The sources output marks the selected source with ^*.

Step 3: Run Your Own NTP Server for the LAN
In larger networks, let one or two internal servers synchronize upstream and serve time to everyone else. Add an allow directive for your subnet to the server’s chrony.conf:
allow 192.168.1.0/24
local stratum 10
The local line lets this server keep serving its (drifting) clock even if it temporarily loses upstream connectivity, which is usually what you want inside a LAN. Open the firewall and restart:
firewall-cmd --permanent --add-service=ntp
firewall-cmd --reload
systemctl restart chronyd
Then point clients at it by replacing their pool lines with:
server ntp1.example.local iburst
Step 4: Verify Synchronization Across the Fleet
Do not trust a configuration you have not measured. These commands give you the full picture when you configure Chrony on RHEL 9:
timedatectl
chronyc tracking
chronyc sourcestats
timedatectl should report System clock synchronized: yes and NTP service: active. Accurate time also keeps monitoring data trustworthy — a Zabbix or similar server with a skewed clock graphs nonsense, as anyone who has installed a monitoring server quickly learns.
Step 5: Force a Quick Correction on a Badly Drifted Clock
If a machine has been off for months (a restored VM snapshot, for example), its clock may be far outside the range Chrony slews gently. Force a step correction:
systemctl stop chronyd
chronyd -q "pool vn.pool.ntp.org iburst"
systemctl start chronyd
Alternatively, run chronyc makestep against the running daemon if your makestep policy allows it. Avoid stepping clocks on database servers during peak writes; schedule it in a maintenance window.
Step 6: Keep an Eye on Chrony with Logs and Measurements
Chrony can record how well it is performing, which is invaluable when someone claims “the server clock was wrong last Tuesday”. Enable measurement and statistics logging in chrony.conf:
logdir /var/log/chrony
log measurements statistics tracking
Restart the service and the files in /var/log/chrony/ will show offsets and frequency corrections over time. Even without logging, make chronyc tracking part of your regular health checks; a sudden jump in the System time offset is often the first symptom of a dying CMOS battery or a broken upstream source.
Troubleshooting
chronyc tracking shows “Leap status: Not synchronised”: no source is reachable. Check UDP 123 in the firewall, verify DNS resolution of your pool names, and inspect chronyc sources for ^? (unreachable) markers.
Sources show huge offsets in a VM: disable competing time-sync services from the hypervisor (for example the VMware or Hyper-V time service) so only Chrony disciplines the clock.
timedatectl says NTP is inactive: another service such as systemd-timesyncd (on some rebuilds) may be conflicting, or chronyd failed to start — read journalctl -u chronyd for the reason.
Conclusion
To configure Chrony on RHEL 9 you only need three habits: point clients at sensible sources with iburst, serve your LAN from one or two internal servers with an allow rule, and verify with chronyc tracking instead of assuming. Do that, and an entire category of mysterious certificate, authentication and logging bugs simply never happens. The official Chrony project site documents every directive when you need advanced setups.