Skip to content

Troubleshooting

Common FreeIPA errors and their solutions : DNS failures, Kerberos clock skew, SSSD issues, backup/restore, and production hardening.

Artikel ini belum diterjemahkan. Isi berikut berbahasa Inggris. Buka versi asli bahasa Inggris.

Troubleshooting

This page covers the most common FreeIPA deployment issues, backup and recovery, and production hardening advice.

Common Deployment Errors

Error: DNS Resolution Failed

Symptom:

ipa-server-install: ERROR: DNS resolution of ipa-server.example.internal failed

Diagnosis:

cat /etc/hosts | grep ipa-server
getent hosts ipa-server.example.internal
cat /etc/resolv.conf

Fix: Ensure /etc/hosts has the correct entry with FQDN first:

echo "192.168.100.128 ipa-server.example.internal ipa-server" >> /etc/hosts

Error: NTP Synchronization Failed

Symptom:

ipa-server-install: ERROR: NTP synchronization failed

Diagnosis:

systemctl status chronyd
chronyc sources
chronyc makestep
timedatectl status

Fix: If the NTP source is unreachable or blocked by firewall:

systemctl stop chronyd
ntpdate pool.ntp.org
systemctl start chronyd

Root cause: NTP source unreachable or UDP port 123 blocked.

Error: Kerberos Key Table Already Exists

Symptom:

ipa-server-install: ERROR: The Kerberos key table already exists

Fix:

ipa-server-install --uninstall
rm -f /etc/krb5.keytab
# Then re-run ipa-server-install

Error: Directory Server Failed to Start

Symptom:

ipa-server-install: ERROR: Directory Server failed to start

Diagnosis:

journalctl -u dirsrv -n 50
df -h /var
ss -tlnp | grep 389
getenforce

Common causes: Disk full, port 389 already in use, or overly restrictive SELinux policies.

Client-Side Issues

SSH Login Timeout

Symptom: ssh testuser@client-1 hangs and eventually times out.

Diagnosis:

systemctl status sssd
journalctl -u sssd -n 100
ping ipa-server.example.internal
dig +short ipa-server.example.internal
cat /etc/krb5.conf
kinit -V testuser

Common fixes:

Cause Solution
Network/DNS unreachable Check firewall, DNS config
SSSD not running systemctl start sssd
SELinux blocking getenforce, then setenforce 0 to test
Misconfigured Kerberos Verify [realm] and [domain_realm] in /etc/krb5.conf

User Not Found on Client

Symptom:

id testuser
# id: 'testuser': no such user

Diagnosis:

# Does the user exist on the server?
ipa user-find testuser

# Clear the SSSD cache
sss_cache -u testuser

# Check NSS configuration
grep sss /etc/nsswitch.conf

# Manual LDAP lookup
ldapsearch -x -h ipa-server.example.internal -b "cn=users,cn=accounts,dc=example,dc=internal" uid=testuser

Common causes: User not created, SSSD not synced, sss missing from /etc/nsswitch.conf, or LDAP connection blocked.

Password Change Not Recognized

Symptom: After ipa passwd testuser, the new password fails.

Fix:

kdestroy -A                  # Clear Kerberos tickets
sss_cache -u testuser        # Clear SSSD cache
sleep 30                     # Wait for SSSD sync
ssh testuser@client-1        # Retry

Root cause: SSSD caches credentials. Changes need cache flush or sync delay to propagate.

Kerberos Clock Skew

Symptom:

kinit testuser
# KDC returned error string: SKEW(Clock skew too great)

Diagnosis:

timedatectl status
date
chronyc sources

Fix:

chronyc makestep                    # Force NTP sync
ntpdate -s pool.ntp.org             # Fallback manual sync

# If the server and client times differ:
ssh ipa-server 'date'
date --set="$(ssh ipa-server date)" # Match server time manually

Root cause: NTP misconfigured or not running. Kerberos requires time difference under 5 minutes.

Backup and Recovery

Create a Backup

ipa-backup
# Creates: /var/lib/ipa/backup/ipa-full-YYYY-MM-DD-HH-MM-SS.tar.gz

What gets backed up: LDAP database, Kerberos keys, PKI certificates, DNS zone data, and configuration files.

Schedule regular backups:

0 2 * * 0 /usr/sbin/ipa-backup

Restore from Backup

ls -la /var/lib/ipa/backup/
ipa-restore /var/lib/ipa/backup/ipa-full-YYYY-MM-DD-HH-MM-SS.tar.gz
ipactl status

Uninstalling FreeIPA

Uninstall a Client

ipa-client-install --uninstall

Uninstall the Server

# WARNING: This destroys all FreeIPA data
ipa-server-install --uninstall

# Remove residual files (only if you are sure)
rm -rf /etc/ipa /var/lib/ipa

Production Hardening

High Availability

FreeIPA supports master-replica architecture for LDAP data replication. Set up a second server as a replica:

  • DNS should resolve to a load-balanced address (e.g., ipa.example.internal pointing to both IPs)
  • Each replica independently runs DNS, Kerberos, and CA services
  • Reference: FreeIPA Replica Setup

Security Hardening

  • Use SSH key authentication instead of passwords
  • Store admin and DS passwords in a secure vault
  • Enable DNSSEC validation (remove --no-dnssec-validation)
  • Keep SELinux enabled, adjust policies instead of disabling

Performance Tuning

  • LDAP cache: SSSD caches by default for 1–2 hours; adjust based on user change frequency
  • DNS cache: Increase TTL (e.g., 3600 seconds) for stable environments
  • Connection pooling: FreeIPA manages this automatically

Systematic Debugging

When troubleshooting, follow this order:

DNS → Network → SSSD → LDAP → Kerberos

Most FreeIPA failures trace back to a broken link somewhere in this chain. Start at the top and work down.

Next Step


Adapted from Dawn’s Blog : FreeIPA Complete Deployment Guide.

Ketik untuk mencari…

↑↓ navigasi↵ pilihEsc tutup