Jtdcftul

I've installed LDAP, BIND9 and Bind9-dyndb-ldap on 18.04 server. Added the BIND schema to LDAP; added a zone and a few server dns entries to LDAP. Added the "dyndb" section to the named.conf, configuring the ldap connectivity information.

...

dyndb "bind9-dyndb" "/usr/lib/bind/ldap.so" {

        uri         "ldap://ldap.example.com";

        base        "ou=dns,dc=example,dc=com";

        auth_method "simple";

        bind_dn     "cn=admin,dc=example,dc=com";

        password    "MyPassword";

};

...

I can search my ldap using the username and password given using "ldapsearch". I also was able to work around the AppArmor config issue that did not allow BIND9 use "/usr/lib/bind/ldap.so".

Now when I attempt to start bind9, I receive the following error in the syslog:

Dec 21 21:42:20 ltserver1 named[17111]: cannot parse settings for 'named.conf for database bind9-dyndb': not found

Dec 21 21:42:20 ltserver1 named[17111]: dynamic database 'bind9-dyndb' configuration failed: not found

Dec 21 21:42:20 ltserver1 named[17111]: loading configuration: not found

Dec 21 21:42:20 ltserver1 named[17111]: exiting (due to fatal error)

I'm stumped as to how to debug this further. Any "pointers" as to what to do would be greatly appreciated!

I have not found any tutorials showing recent bind9-dyndb-ldap/BIND9 configuration examples (the most recent one I find is for Ubuntu 16.04 and it still uses "Dynamic-DB" instead of "dyndb"). I feel like I'm groping in the dark. Any assistance would be very helpful!

UPDATE: SOLVED
I found this Bugzilla link at Red Hat: https://bugzilla.redhat.com/show_bug.cgi?id=1436268

If server_id is not defined, it crashes.

So, I just added

server_id "[hostname of the machine]"

to the dyndb section of my named.conf file, and it started right up.

Hope this helps.

Original Post

I'm upgrading from Xenial to Bionic and have the exact same problem, and an almost identical config as yours. Best I can tell, I've rewritten my config file into the new dyndb format. Had to change the option formats from:

arg "uri ldaps://example.com"

to

uri "ldaps://example.com"

The only difference is my password is unquoted as the directive is for an unquoted string (though it doesn't complain if quoted).

No matter what debug level I set, I don't seem to get any more details that it cannot parse something. All of those are valid config directives, so I'm stumped!

Since this is all being built in docker, I can always use the Xenial image as my older DNS container images build and deploy fine with the older packages, but I'd like to figure out what part of the config this one doesn't like.

For what it's worth, I did get it to spit out the expected grammar for the dyndb LDAP module, though it was all what I expected.

22-Jan-2019 01:29:03.990 expected grammar:

auth_method <quoted_string>;

base <quoted_string>;

bind_dn <quoted_string>;

connections <integer>;

directory <quoted_string>;

dyn_update <boolean>;

fake_mname <quoted_string>;

krb5_keytab <quoted_string>;

krb5_principal <quoted_string>;

ldap_hostname <quoted_string>;

password <string>;

reconnect_interval <integer>;

sasl_auth_name <quoted_string>;

sasl_mech <quoted_string>;

sasl_password <quoted_string>;

sasl_realm <quoted_string>;

sasl_user <quoted_string>;

server_id <quoted_string>;

sync_ptr <boolean>;

timeout <integer>;

uri <quoted_string>;

verbose_checks <boolean>;