During an upgrade from 4.5.0-21.el7.centos.1.2
to 4.5.0-21.el7.centos.2.2 on a CentOS 7.4 machine, FreeIPA's DNS server
briefly returned NXDOMAIN for records which existed in FreeIPA. These
invalid responses were returned for a very short amount of time, but
caused long-running issues with Java clients which tend to cache DNS
responses. Upgraded packages included: 389-ds-base, 389-ds-base-libs,
389-ds-base-snmp, ipa-client, ipa-client-common, ipa-python-compat,
ipa-server, ipa-server-common, ipa-server-dns, ipa-server-trust-ad,
python2-ipa-server, and a dozen sss-related packages.
I reproduced this in a FreeIPA test environment by running `while true;
do dig some.dns.entry.managed.by.freeipa @ip.address.of.freeipa | tee -a
a-log-file; done` from one server, and running `yum update` on the
FreeIPA machine. The invalid NXDOMAIN responses were returned some time
after the `yum update` logged 'Cleanup' for the RPMs, and seemed to be
during the 'Verifying' phase.
These NXDOMAIN responses claimed that an upstream nameserver
(
a.root-servers.net <
http://a.root-servers.net>) was the authority for
my zone:
a-log-file-; <<>> DiG 9.8.2rc1-RedHat-9.8.2-0.37.rc1.el6_7.7
<<>>
some.dns.entry.managed.by.freeipa @172.16.0.77 <
http://172.16.0.77>
a-log-file-;; global options: +cmd
a-log-file-;; Got answer:
a-log-file:;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 2889
a-log-file-;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1,
ADDITIONAL: 0
a-log-file-
a-log-file-;; QUESTION SECTION:
a-log-file-;some.dns.entry.managed.by.freeipa. IN A
a-log-file-
a-log-file-;; AUTHORITY SECTION:
a-log-file-.60INSOAa.root-servers.net <
http://a.root-servers.net>.
nstld.verisign-grs.com <
http://nstld.verisign-grs.com>. 2017102400 1800
900 604800 86400
a-log-file-
a-log-file-;; Query time: 227 msec
a-log-file-;; SERVER: 172.16.0.77#53(172.16.0.77)
a-log-file-;; WHEN: Tue Oct 24 18:30:28 2017
a-log-file-;; MSG SIZE rcvd: 130
Usually when querying an invalid DNS entry, the dig output still claims
that my FreeIPA server is authoritative for the zone:
$ dig doesntexist.zone.managed.by.freeipa @172.16.0.77 <
http://172.16.0.77>
; <<>> DiG 9.8.2rc1-RedHat-9.8.2-0.37.rc1.el6_7.7 <<>>
doesntexist.zone.managed.by.freeipa @172.16.0.77 <
http://172.16.0.77>
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 59953
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
;; QUESTION SECTION:
;doesntexist.zone.managed.by.freeipa. IN A
;; AUTHORITY SECTION:
zone.managed.by.freeipa.30 INSOAidm01.freeipa.
hostmaster.zone.managed.by.freeipa. 1508869828 30 900 1209600 30
;; Query time: 0 msec
;; SERVER: 172.16.0.77#53(172.16.0.77)
;; WHEN: Tue Oct 24 19:27:12 2017
;; MSG SIZE rcvd: 113
Is it possible that during a yum update, the FreeIPA DNS server
temporarily forgets what zones it's authoritative for (or forgets all
DNS records) and just delegates to the upstream DNS server for half a
second or so? Or is something else going on here?
I'm open to suggestions.
The LDAP server is brought down during upgrades which is likely the
issue. bind can't connect to its backend. Why it returns NXDOMAIN I
don't know.
You may be able to manually work around this by manually stopping bind
before updating IPA, then starting it again afterwards.
rob