DHCP- keine neuen Informationen aus LDAP

UCS - Univention Corporate Server
, ,
#1

Ich richte gerade DHCP für ein Netzwerk ein.
Es existiert bereits das Subnetz mit 10.98.0.0/16. Bei diesem dhcp-subnetz habe ich keine Range eingetragen.

Dann gibt es 3 Pools in diesem Subnetz. Zwei für statische und eins für dynamische vergabe. Dort habe ich einmal Unbekannte Geräte auf deaktiviert und bekannte auf aktiv und für den dynamischen pool anders herum konfiguriert.
Das sind die Drei Pools:
univention-ldapsearch -b cn=dhcp,$(ucr get ldap/base) ‘(!(objectClass=univentionDhcpHost))’ | grep --color -P "Range"

dhcpRange: 10.98.250.0 10.98.252.254
dhcpRange: 10.98.64.1 10.98.64.254
dhcpRange: 10.98.0.1 10.98.0.254

Jedoch taucht in meiner /var/log/syslog folgende Fehler auf:

...............
Feb  6 10:52:50 ucs-master dhcpd[16003]: LDAP line 24: lease 10.98.0.240 is declared twice!
Feb  6 10:52:50 ucs-master dhcpd[16003]: range 10.98.0.1 10.98.0.254;
Feb  6 10:52:50 ucs-master dhcpd[16003]:                             ^
Feb  6 10:52:50 ucs-master dhcpd[16003]: LDAP line 24: lease 10.98.0.241 is declared twice!
Feb  6 10:52:50 ucs-master dhcpd[16003]: range 10.98.0.1 10.98.0.254;
Feb  6 10:52:50 ucs-master dhcpd[16003]:                             ^

Diese Fehler wiederholen sich auch bei dem Pool mit der Range von 10.98.250.1-10.98.252.254.

Ich habe das Gefühl das der dhcpd nicht die aktuellen Informationen bezieht.
Habe schon folgendes versucht:

  • dhcp neustarten
  • univention server neustarten
  • configuration erneut anlegen

Habe soweit auch schon den Teil: IP-Vergabe über DHCP durchgelesen.

Habt ihr eine Idee?

English summary:

The user has created a DHCP subnet with three DHCP pools in it: two pools for static address assignment and one for dynamic assignments.

Unfortunately this results in the DHCP server not starting anymore and error messages being emitted in the system’s journal/log files.

#2

Hey,

for the benefit of our international readers I’ll reply in English. If that’s a problem for you, just give me a shout and I’ll switch back to German.

This looks like a bug to me and is really easy to reproduce. I haven’t found an entry in the bug tracker yet; therefore I’ve just created one.

Kind regards
mosu

1 Like
#3

Thanks for your reply Moritz,
I’ll changed the lease-time policy to a very short time now, just for debug and somehow after a short break the whole think worked nearly as aspected. Maybe not bug for bad functionality but for lag between changes and the real change for the dhcp daemon to be performed.
I will comment the bug for integrity of the problem.

Thanks for your help. Next thread will be in english!

#4

Hm, strange. Just a remark or for future reference: Something that helped me a lot in the past is to uncomment the following line in /etc/dhcp/dhcpd.conf:

# ldap-debug-file "/var/log/dhcp-ldap-startup.log";

Then restarting the DHCP service and having a look at /var/log/dhcp-ldap-startup.log which should show the actual configuration the DHCP daemon read from the LDAP directory.

Best regards,
Michael Grandjean

1 Like
#5

After some experience with Univention and networking i found my own post here at help.univention.com

To all beginners which see such console errors: Declared twice is basicly the thing that stares into my eyes when i see such errors.

Go to your ucs -> domain -> dhcp
look for the dhcp service where the ips given in the log are managed.
examine all pools! If one or more of them are overlapping with other pools dhcp wont start.
@Grandjean would you consider this an bug or a feature?

best regards!

Mastodon