With C202 systems on my LAN with DHCP assigned addresses could talk to each other by name. With C203, this is no longer the case.
Based on my understanding of various topics I’ve read here, I have disabled and enabled DHCP, rebooted the IPFire server, and rebooted the systems on the LAN that are having problems.
I’m not sure what additional information I need to provide to get this resolved.
Use edit host now.
Up to CU202 the DHCP page of WUI generated default host names using the remarks for fixed leases.
This feature is not implemented in CU203, anymore. If the local devices do not supply a host name in their DHCP responses, the WUI does not send one to the DNS server/resolver.
As @hvacguy mentioned you have to give this information in the hosts definition.
I’ve done everything suggested in the Core Update 203: Recommendations Before and After Upgrading topic.
The current DHCP configuration:
Below are dig results for working brad-morgan5.morgan.local (fixed lease, manually added to hosts) and a failing brad-morgan6.morgan.local (DHCP assigned address).
Not sure what else to add.
[root@bricknix ~]# dig brad-morgan5.morgan.local
; <<>> DiG 9.20.23 <<>> brad-morgan5.morgan.local
;; global options: +cmd
;; Got answer:
;; WARNING: .local is reserved for Multicast DNS
;; You are currently testing what happens when an mDNS query is leaked to DNS
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 7628
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;brad-morgan5.morgan.local. IN A
;; ANSWER SECTION:
brad-morgan5.morgan.local. 60 IN A 192.168.0.163
;; Query time: 0 msec
;; SERVER: 127.0.0.1#53(127.0.0.1) (UDP)
;; WHEN: Mon Aug 31 06:13:05 MDT 2026
;; MSG SIZE rcvd: 70
[root@bricknix ~]# dig brad-morgan6.morgan.local
; <<>> DiG 9.20.23 <<>> brad-morgan6.morgan.local
;; global options: +cmd
;; Got answer:
;; WARNING: .local is reserved for Multicast DNS
;; You are currently testing what happens when an mDNS query is leaked to DNS
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 23789
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;brad-morgan6.morgan.local. IN A
;; AUTHORITY SECTION:
local. 3600 IN SOA localhost. nobody.invalid. 1 3600 1200 604800 10800
;; Query time: 0 msec
;; SERVER: 127.0.0.1#53(127.0.0.1) (UDP)
;; WHEN: Mon Aug 31 06:13:16 MDT 2026
;; MSG SIZE rcvd: 113
[root@bricknix ~]#
brad-morgan5 has an IP out the set of dynamic leases. This is not recommended, see my various annotations about this case.
How is brad-morgan6 defined?
The propagation of (host name, IP) tuples from DHCP to DNS is documented in the DHCP section of /var/log/messages (Logs–>System Logs menu of WUI). What informations are given there for the morgan6 device?
Please check the logs to make sure that your client brad-morgan6 is actually sending this hostname when connecting and that it is being assigned the correct IP address:
tail -f /var/log/messages | grep dhcp
You should see something similar to:
Aug 31 14:35:44 bricknix dhcpd: execute_statement argv[0] = /usr/sbin/dhcp-lease
Aug 31 14:35:44 bricknix dhcpd: execute_statement argv[1] = commit
Aug 31 14:35:44 bricknix dhcpd: execute_statement argv[2] = ADDRESS=192.168.0.xx
Aug 31 14:35:44 bricknix dhcpd: execute_statement argv[3] = HOSTNAME=brad-morgan6
Aug 31 14:35:44 bricknix dhcpd: execute_statement argv[4] = DOMAINNAME=morgan.local
Aug 31 14:35:44 bricknix dhcpd: DHCPREQUEST for 192.168.0.xx from xx:xx:xx:xx:xx:xx via green0
Aug 31 14:35:44 bricknix dhcpd: DHCPACK on 192.168.0.xx to xx:xx:xx:xx:xx:xx via green0
If the client is sending a different hostname, you may also need to declare it in hosts.cgi.
Finally, please also check the contents of the leases.db database to make sure it is being populated correctly:
sqlite3 /var/lib/knot-resolver/leases.db <<<"select * from leases"
With C202, brad-morgan5 was DHCP assigned and hosts a SageTV web server. I have an iPad next to my TV chair which accesses the program guide provided by SageTV.
When I upgraded to C203, the iPad quit working. As a temporary measure, I added the fixed lease and added brad-morgan5 to hosts so I could watch/record broadcast TV as I have been for many years.
I think this is what you asked for:
Aug 31 07:07:33 bricknix dhcpd: execute_statement argv[0] = /usr/sbin/dhcp-lease
Aug 31 07:07:33 bricknix dhcpd: execute_statement argv[1] = commit
Aug 31 07:07:33 bricknix dhcpd: execute_statement argv[2] = ADDRESS=192.168.0.107
Aug 31 07:07:33 bricknix dhcpd: execute_statement argv[3] = HOSTNAME=Brad-Morgan6
Aug 31 07:07:33 bricknix dhcpd: execute_statement argv[4] = DOMAINNAME=morgan.local
Aug 31 07:07:34 bricknix dhcpd: DHCPREQUEST for 192.168.0.107 from 24:4b:fe:58:95:59 (Brad-Morgan6) via green0
Aug 31 07:07:34 bricknix dhcpd: DHCPACK on 192.168.0.107 to 24:4b:fe:58:95:59 (Brad-Morgan6) via green0
Aug 31 07:07:43 bricknix dhcpd: DDNS: cleaning up lease pointer for a cancel cb=0x2a37d6e0
Aug 31 07:07:43 bricknix dhcpd: Unable to add forward map from Brad-s-Tab-S8.morgan.local to 192.168.0.161: operation canceled
Aug 31 07:07:46 bricknix dhcpd: DDNS: cleaning up lease pointer for a cancel cb=0x2a32f8f0
Aug 31 07:07:46 bricknix dhcpd: Unable to add forward map from Brad-Morgan6.morgan.local to 192.168.0.107: operation canceled
[root@bricknix ~]# sqlite3 /var/lib/knot-resolver/leases.db <<<"select * from leases"
╭───────────────┬───────────────────────────────────╮
│ address │ hostname │
╞═══════════════╪═══════════════════════════════════╡
│ 192.168.0.161 │ brad-s-tab-s8.morgan.local. │
│ 192.168.0.162 │ eero.morgan.local. │
│ 192.168.0.176 │ esp_7d23c6.morgan.local. │
│ 192.168.0.118 │ google-nest-mini.morgan.local. │
│ 192.168.0.113 │ esp_7d27df.morgan.local. │
│ 192.168.0.110 │ esp_7d289a.morgan.local. │
│ 192.168.0.158 │ brad-morgan8.morgan.local. │
│ 192.168.0.143 │ myq-5f9.morgan.local. │
│ 192.168.0.170 │ esp_7d249d.morgan.local. │
│ 192.168.0.188 │ myq-dff.morgan.local. │
│ 192.168.0.172 │ brad-morgan7.morgan.local. │
│ 192.168.0.134 │ aicwlan.morgan.local. │
│ 192.168.0.107 │ brad-morgan6.morgan.local. │
│ 192.168.0.109 │ rpi3b.morgan.local. │
│ 192.168.0.191 │ denali-5cfce1fd9184.morgan.local. │
│ 192.168.0.156 │ hs300.morgan.local. │
│ 192.168.0.101 │ hs105.morgan.local. │
│ 192.168.0.148 │ hdhr-10b29a45.morgan.local. │
│ 192.168.0.151 │ libre.morgan.local. │
│ 192.168.0.180 │ rpi0w.morgan.local. │
│ 192.168.0.123 │ octopi.morgan.local. │
│ 192.168.0.182 │ rpi4b.morgan.local. │
│ 192.168.0.192 │ kp303.morgan.local. │
│ 192.168.0.189 │ hdhr-103ceacf.morgan.local. │
│ 192.168.0.190 │ hdhr-103c9e78.morgan.local. │
│ 192.168.0.141 │ brad-morgan2.morgan.local. │
│ 192.168.0.146 │ c110.morgan.local. │
│ 192.168.0.102 │ hs103.morgan.local. │
│ 192.168.0.165 │ c210.morgan.local. │
│ 192.168.0.168 │ brad-morgan3.morgan.local. │
│ 192.168.0.145 │ ipad.morgan.local. │
│ 192.168.0.185 │ brad-s-s22.morgan.local. │
│ 192.168.0.163 │ brad-morgan5.morgan.local. │
╰───────────────┴───────────────────────────────────╯
[root@bricknix ~]#
The mysteries of CU203…
try
/etc/init.d/knot-resolver restart
Wait a few seconds for the resolver to restart, then run:
kresctl cache clear
The results:
[root@bricknix ~]# /etc/init.d/knot-resolver restart
Stopping Knot Resolver Daemon... [ OK ]
Starting Knot Resolver Daemon... [ OK ]
[root@bricknix ~]# kresctl cache clear
count: 22733
[root@bricknix ~]# dig brad-morgan6.morgan.local
; <<>> DiG 9.20.23 <<>> brad-morgan6.morgan.local
;; global options: +cmd
;; Got answer:
;; WARNING: .local is reserved for Multicast DNS
;; You are currently testing what happens when an mDNS query is leaked to DNS
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 21337
;; flags: qr rd ra ad; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 1232
;; QUESTION SECTION:
;brad-morgan6.morgan.local. IN A
;; AUTHORITY SECTION:
local. 3600 IN SOA localhost. nobody.invalid. 1 3600 1200 604800 10800
;; Query time: 0 msec
;; SERVER: 127.0.0.1#53(127.0.0.1) (UDP)
;; WHEN: Mon Aug 31 07:26:41 MDT 2026
;; MSG SIZE rcvd: 113
[root@bricknix ~]#
Probable issue
Have you tried adding brad-morgan6 with the IP address 192.168.0.107 in hosts.cgi?
I really don’t have any other ideas at this point.
That’s what I did to solve the “brad-morgan5” problem. I haven’t (yet) found any issues with brad-morgan6 so I’ll leave it alone for now.
#13973 seems to be solved, really. The host name is syntactically right to DNS.
But the automagically generated host name of CU202 and former is not implemented in CU203.
The latter is know widely now. We have discussed it several times.
Unchecking “Enable DNS Update (RFC2136)” appears to have resolved this issue. It’s not immediately obvious from the wording of that option so maybe the text could be clarified.
Thanks to all who helped figure this out.
I think the “DNS Update” is clarified enough ( for those who want to use it ) by referencing the RFC and the description in the wiki. Latter should updated to knot functionality, thats true. But if you choose the fact ‘internal resolver of IPFire does not support RFC2136’ you are on the safe side.
It’s been corrected.
