Netgate Discussion Forum
    • Categories
    • Recent
    • Tags
    • Popular
    • Users
    • Search
    • Register
    • Login

    2.1-BETA0 -> 2.1-BETA1 unbound won't install

    Scheduled Pinned Locked Moved 2.1 Snapshot Feedback and Problems - RETIRED
    63 Posts 14 Posters 20.8k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • rbgargaR
      rbgarga Developer Netgate Administrator
      last edited by

      Please try 1.4.20_4 and let me know the results.

      Renato Botelho

      1 Reply Last reply Reply Quote 0
      • A
        athurdent
        last edited by

        Looks really good on the commandline now, only one unbound_monitor.sh. Works fine!
        But it still restarts every hour, any idea what might be causing that?

        1 Reply Last reply Reply Quote 0
        • rbgargaR
          rbgarga Developer Netgate Administrator
          last edited by

          @athurdent:

          Looks really good on the commandline now, only one unbound_monitor.sh. Works fine!
          But it still restarts every hour, any idea what might be causing that?

          No idea until now, I couldn't replicate it on my local test env.

          Renato Botelho

          1 Reply Last reply Reply Quote 0
          • J
            jcyr
            last edited by

            Not quite there yet! Every time I change and save new dhcpd settings a new unbound_monitor.sh is added.

            IPV6 Test: http://ipv6-test.com

            1 Reply Last reply Reply Quote 0
            • J
              jcyr
              last edited by

              At the beginning of the /usr/local/etc/rc.d/unbound_monitor.sh script, there is the following:

              **PROCS=/bin/pgrep -f unbound_monitor.sh | wc -l | awk '{print $1}'

              if [ ${PROCS} -gt [color=red]2 ]; then
                     echo "There are another unbound monitor proccess running"
                     exit 0
              fi**
              Shouldn't it be:

              **PROCS=/bin/pgrep -f unbound_monitor.sh | wc -l | awk '{print $1}'

              if [ ${PROCS} -gt [color=red]1 ]; then
                     echo "There are other unbound monitor proccess running"
                     exit 0
              fi**
              ???

              IPV6 Test: http://ipv6-test.com

              1 Reply Last reply Reply Quote 0
              • A
                athurdent
                last edited by

                Still reloads every hour, I cannot find a reason for that. The only thing I found out was that I have to issue```
                unbound-control reload

                
                Another problem is that my DMZ on opt1 is missing the option```
                domain-name-servers 192.168.1.254;
                ```in /var/dhcpd/etc/dhcpd.conf when I use unbound. I have opt1 defined as Network interface in unbound configuration and I can "dig" it from the DMZ. Using the system forwarder the option is there.
                1 Reply Last reply Reply Quote 0
                • rbgargaR
                  rbgarga Developer Netgate Administrator
                  last edited by

                  unbound_monitor.sh was not working with last change, I changed the logic a bit and the way it's called and released 1.4.20_5. Let me know if you find any issues. I tested it on nanobsd and full installation.

                  Renato Botelho

                  1 Reply Last reply Reply Quote 0
                  • A
                    athurdent
                    last edited by

                    I found the reason for the unbound reloads on my syslog server. Why would dhcpleases do this?

                    May  1 07:29:20 pfsense dhcpleases: Sending HUP signal to dns daemon(36838)
                    May  1 07:29:20 pfsense unbound: [36838:0] info: service stopped (unbound 1.4.20).
                    

                    And domain-name-servers is still missing from my opt1 DHCPD config as stated in my reply above.
                    Other than that, everything looks good now, only one monitor process.
                    Thanks for looking into this!

                    1 Reply Last reply Reply Quote 0
                    • W
                      wallabybob
                      last edited by

                      @athurdent:

                      I found the reason for the unbound reloads on my syslog server. Why would dhcpleases do this?

                      PERHAPS it sends a signal to the local DNS to tell it the local name to IP address mapping has changed and the local DNS should read the updated information.

                      1 Reply Last reply Reply Quote 0
                      • A
                        athurdent
                        last edited by

                        Hmm, but I have not enabled "Register DHCP static mappings". What else could cause this?

                        1 Reply Last reply Reply Quote 0
                        • W
                          wallabybob
                          last edited by

                          @athurdent:

                          Hmm, but I have not enabled "Register DHCP static mappings". What else could cause this?

                          Perhaps it is trying to register a DHCP dynamic mapping! The thread at http://forum.pfsense.org/index.php/topic,33240.msg suggests dhcpleases sends a signal to the DNS to update the DNS registrations.

                          DNS forwarder has an option Register DHCP leases in DNS forwarder. Do you have the equivalent set in Unbound? I don't know about Unbound and don't run it - does it "inherit" DNS forwarder settings?

                          Maybe Unbound shouldn't stop when it gets a signal from dhcpleases.

                          1 Reply Last reply Reply Quote 0
                          • A
                            athurdent
                            last edited by

                            Thanks wallabybob, now I was able to solve it:
                            I had to turn DNS Forwarder back on, uncheck
                            "Register DHCP leases in DNS forwarder"
                            turn it off again and finally reboot. Otherwise dhcpleases would not pick up the change.
                            It (or something else) filled
                            /var/etc/hosts
                            with many of these:

                            
                            # dhcpleases automatically entered
                            
                            # dhcpleases automatically entered
                            
                            # dhcpleases automatically entered
                            
                            

                            Now they don't show up anymore an unbound runs uninterrupted.

                            Thanks again for the help!

                            1 Reply Last reply Reply Quote 0
                            • rbgargaR
                              rbgarga Developer Netgate Administrator
                              last edited by

                              @athurdent:

                              I found the reason for the unbound reloads on my syslog server. Why would dhcpleases do this?

                              May  1 07:29:20 pfsense dhcpleases: Sending HUP signal to dns daemon(36838)
                              May  1 07:29:20 pfsense unbound: [36838:0] info: service stopped (unbound 1.4.20).
                              

                              And domain-name-servers is still missing from my opt1 DHCPD config as stated in my reply above.
                              Other than that, everything looks good now, only one monitor process.
                              Thanks for looking into this!

                              Just to make sure, does this issue still persist? If yes, could you send more details about it?

                              Renato Botelho

                              1 Reply Last reply Reply Quote 0
                              • A
                                athurdent
                                last edited by

                                Hi Renato,

                                thanks for asking!

                                the reload issue was tracked down, please see post #55 for the details. Maybe something could be done about it by clearing the system DNS forwarder's config when it's diabled? Or by making "Register DHCP leases in DNS forwarder" an option in unbound an syncing it with the system setting?

                                The "domain-name-servers" entry still does not get configured for my opt1 as it does when using the system DNS forwarder. As a workaround I configured this entry in my DHCP config for opt1 now. What details do you need? I guess it can be easily reproduced by configuring an opt1 interface, enabling DHCP Server/Unbound for that interface and looking at /var/dhcpd/etc/dhcpd.conf. It should be missing the domain-name-servers entry for the opt1 IP.

                                1 Reply Last reply Reply Quote 0
                                • rbgargaR
                                  rbgarga Developer Netgate Administrator
                                  last edited by

                                  @athurdent:

                                  the reload issue was tracked down, please see post #55 for the details. Maybe something could be done about it by clearing the system DNS forwarder's config when it's diabled? Or by making "Register DHCP leases in DNS forwarder" an option in unbound an syncing it with the system setting?

                                  I pushed a fix for this, it was not being checked if DNS Fowarder was enabled or not. It should be fine on current 2.1 code.

                                  @athurdent:

                                  The "domain-name-servers" entry still does not get configured for my opt1 as it does when using the system DNS forwarder. As a workaround I configured this entry in my DHCP config for opt1 now. What details do you need? I guess it can be easily reproduced by configuring an opt1 interface, enabling DHCP Server/Unbound for that interface and looking at /var/dhcpd/etc/dhcpd.conf. It should be missing the domain-name-servers entry for the opt1 IP.

                                  It's done automatically for DNS Forwarder but not for unbound, the only way is to configure it on DHCP config, as you already did.

                                  Renato Botelho

                                  1 Reply Last reply Reply Quote 0
                                  • M
                                    mromero
                                    last edited by

                                    Hi everyone, does this mean Unbound has now been fixed to work with PfSense 2.1?

                                    1 Reply Last reply Reply Quote 0
                                    • rbgargaR
                                      rbgarga Developer Netgate Administrator
                                      last edited by

                                      @mromero:

                                      Hi everyone, does this mean Unbound has now been fixed to work with PfSense 2.1?

                                      Yes.

                                      Renato Botelho

                                      1 Reply Last reply Reply Quote 0
                                      • J
                                        jcyr
                                        last edited by

                                        Thanks Renato

                                        IPV6 Test: http://ipv6-test.com

                                        1 Reply Last reply Reply Quote 0
                                        • A
                                          athurdent
                                          last edited by

                                          Could we have an option to turn off IPv6 support?
                                          pfSense does not handle IPv6 DNSSEC queries if they get fragmented (i.e. when using a tunnel broker for IPv6 access):

                                          May 24 13:54:27 vpn pf: 00:01:05.283725 rule 5/0(match): block in on gif0: (hlim 59, next-header Fragment (44) payload length: 1240) 2001:500:1b::1 > 2001:xxxx:xxxx:xxxx::2: frag (0x82af3ff7:0|1232) 53 > 39396: 34110*- 7/0/1 info. DNSKEY[|domain]
                                          May 24 13:54:27 vpn pf: 00:00:00.000123 rule 5/0(match): block in on gif0: (hlim 59, next-header Fragment (44) payload length: 413) 2001:500:1b::1 > 2001:xxxx:xxxx:xxxx::2: frag (0x82af3ff7:1232|405)
                                          

                                          btw, the first line looks odd in when you view it in the web gui. It only shows 53 and 39396 as source and destination, not the IPs.

                                          1 Reply Last reply Reply Quote 0
                                          • First post
                                            Last post
                                          Copyright 2025 Rubicon Communications LLC (Netgate). All rights reserved.