• DNS funny with Debian

    From Joe@3:633/10 to All on Tuesday, July 14, 2026 23:20:01

    Not an urgent one, just something to be solved.

    My old Microserver died. I had been slowly preparing two Raspberry Pis
    to take over, but my hand was forced. I got the main server running
    postfix working reasonably well. To save time, I used the ISP's router
    for DNS and DHCP, intending to replace them later. After a crash course
    in postfix and nftables it all seem to be working.

    Then I switched DNS to BIND9 running on the other Pi. All OK, then an
    email from this list came in and was bounced for having no PTR. I very
    quickly restored DNS to the router and investigated. It turns out that
    the router DNS server, whatever it is, can see bendel, but BIND9
    cannot. I've run bind for decades, the last thirteen years on the
    Microserver, and never had a problem with bendel. The current BIND9 is
    running without forwarders, as previous versions have. Here's what dig
    shows:

    ISP DNS:

    joe@jrenewsid:~$ dig -x 82.195.75.100

    ; <<>> DiG 9.20.24-1+b1-Debian <<>> -x 82.195.75.100
    ;; global options: +cmd
    ;; Got answer:
    ;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 19280
    ;; flags: qr rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 0, ADDITIONAL: 1

    ;; OPT PSEUDOSECTION:
    ; EDNS: version: 0, flags:; udp: 512
    ;; QUESTION SECTION:
    ;100.75.195.82.in-addr.arpa. IN PTR

    ;; ANSWER SECTION:
    100.75.195.82.in-addr.arpa. 851 IN CNAME
    100.64-26.75.195.82.in-addr.arpa.
    100.64-26.75.195.82.in-addr.arpa. 851 IN PTR bendel.debian.org.

    ;; Query time: 8 msec
    ;; SERVER: 192.168.71.1#53(192.168.71.1) (UDP)
    ;; WHEN: Tue Jul 14 21:21:04 BST 2026
    ;; MSG SIZE rcvd: 110

    BIND9:

    joe@jrenewsid:~$ dig @192.168.71.8 -x 82.195.75.100

    ; <<>> DiG 9.20.24-1+b1-Debian <<>> @192.168.71.8 -x 82.195.75.100
    ; (1 server found)
    ;; global options: +cmd
    ;; Got answer:
    ;; ->>HEADER<<- opcode: QUERY, status: SERVFAIL, id: 31340
    ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 1

    ;; OPT PSEUDOSECTION:
    ; EDNS: version: 0, flags:; udp: 1232
    ; COOKIE: 52c91b299124cdfd010000006a569a4c89ea67199c61c1fe (good)
    ;; QUESTION SECTION:
    ;100.75.195.82.in-addr.arpa. IN PTR

    ;; Query time: 96 msec
    ;; SERVER: 192.168.71.8#53(192.168.71.8) (UDP)
    ;; WHEN: Tue Jul 14 21:21:32 BST 2026
    ;; MSG SIZE rcvd: 83

    Also ISP DNS:

    joe@jrenewsid:~$ host 82.195.75.100
    100.75.195.82.in-addr.arpa is an alias for100.64-26.75.195.82.in-addr.arpa. 100.64-26.75.195.82.in-addr.arpa domain name pointer bendel.debian.org.

    Also BIND9:

    joe@piserve1:~ $ host 82.195.75.100
    Host 100.75.195.82.in-addr.arpa not found: 2(SERVFAIL)


    OK, there's a CNAME involved, but should this really stop BIND9 dead?
    Any ideas how to fix this? BIND9 seems to work for everything else I
    tried, and it used to work for bendel on the Microserver, which no
    longer runs.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Greg Wooledge@3:633/10 to All on Tuesday, July 14, 2026 23:40:02
    On Tue, Jul 14, 2026 at 22:14:07 +0100, Joe wrote:
    Also ISP DNS:

    joe@jrenewsid:~$ host 82.195.75.100
    100.75.195.82.in-addr.arpa is an alias for100.64-26.75.195.82.in-addr.arpa. 100.64-26.75.195.82.in-addr.arpa domain name pointer bendel.debian.org.

    Also BIND9:

    joe@piserve1:~ $ host 82.195.75.100
    Host 100.75.195.82.in-addr.arpa not found: 2(SERVFAIL)

    hobbit:~$ host 82.195.75.100
    ;; communications error to 127.0.0.1#53: timed out
    100.75.195.82.in-addr.arpa is an alias for 100.64-26.75.195.82.in-addr.arpa. 100.64-26.75.195.82.in-addr.arpa domain name pointer bendel.debian.org.

    I should've used "time". Oh well. The point is, I got an answer,
    after an initial timeout followed by a retry. I'm using dnscache.

    If both you and I are seeing intermittent failures getting the PTR
    record pointing to bendel, that points to an issue somewhere in the
    DNS infrastructure used by Debian or by (one of?) its providers.

    A deeper investigation might be warranted.

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Wednesday, July 15, 2026 09:10:01
    Hi,

    On Tue, Jul 14, 2026 at 10:14:07PM +0100, Joe wrote:
    ISP DNS:

    joe@jrenewsid:~$ dig -x 82.195.75.100

    [?]

    ;; ANSWER SECTION:
    100.75.195.82.in-addr.arpa. 851 IN CNAME 100.64-26.75.195.82.in-addr.arpa.
    100.64-26.75.195.82.in-addr.arpa. 851 IN PTR bendel.debian.org.

    [?]

    BIND9:

    joe@jrenewsid:~$ dig @192.168.71.8 -x 82.195.75.100

    ; <<>> DiG 9.20.24-1+b1-Debian <<>> @192.168.71.8 -x 82.195.75.100
    ; (1 server found)
    ;; global options: +cmd
    ;; Got answer:
    ;; ->>HEADER<<- opcode: QUERY, status: SERVFAIL, id: 31340
    ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 1

    What does:

    dig +trace -x 82.195.75.100 @192.168.71.8

    output?

    It will show you each step of the recursive resolving.

    Do note that part of this recursion uses DNSSEC. If validation fails for some reason then SERVFAIL is the usual outcome.
    BIND9 as a resolver does do DNSSEC validation by default, whereas your
    ISP router's resolver probably does not. So your DNSSEC validation setup
    may be iffy.

    Anything in bind9's logs?

    $ delv -x 82.195.75.100
    ;; validating 100.64-26.75.195.82.in-addr.arpa/PTR: no valid signature found
    ; fully validated
    100.75.195.82.in-addr.arpa. 3600 IN CNAME 100.64-26.75.195.82.in-addr.arpa.
    100.75.195.82.in-addr.arpa. 3600 IN RRSIG CNAME 13 6 3600 20260725072209 20260711055209 25684 75.195.82.in-addr.arpa. e7FvHXk4Bqz9JLxenWVC4CDMZCj5wsDNfqrAD1Au67FcttQtcJJQGwZi YnyXONXn2HfYMcAvw8wOT4VjMWi0Fw==

    ; unsigned answer
    100.64-26.75.195.82.in-addr.arpa. 28799 IN PTR bendel.debian.org. 100.64-26.75.195.82.in-addr.arpa. 28799 IN RRSIG PTR 13 7 28800 20260730021457 20260620015239 49383 64-26.75.195.82.in-addr.arpa. 4IuODtaiusOd4qJu+Ko48jtvhZs3CDFci6hE/fq7+ECHb/waOrl4zu6R Dhb1O6/mj5UfnuLvaklBUbmEVp8uPQ==

    The above shows that the zone that contains "100.75.195.82.in-addr.arpa"
    is signed and validates, while the zone that contains "100.64-26.75.195.82.in-addr.arpa" is not signed. This is not an ideal
    setup but it should work. If your validator is broken it will choke at
    one of the signed steps.

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marco Moock@3:633/10 to All on Wednesday, July 15, 2026 11:50:01
    Am 14.07.26 um 23:20 schrieb Joe:
    Then I switched DNS to BIND9 running on the other Pi. All OK, then an
    email from this list came in and was bounced for having no PTR. I very quickly restored DNS to the router and investigated. It turns out that
    the router DNS server, whatever it is, can see bendel, but BIND9
    cannot. I've run bind for decades, the last thirteen years on the Microserver, and never had a problem with bendel. The current BIND9 is running without forwarders, as previous versions have. Here's what dig
    shows:

    It is a bug in systemd-resolve, I noticed that and blamed sendmail first...

    I already filed it:
    https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1133842

    --
    Gruá
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marco Moock@3:633/10 to All on Wednesday, July 15, 2026 12:00:01
    Am 14.07.26 um 23:20 schrieb Joe:
    Then I switched DNS to BIND9 running on the other Pi.

    Did you disable systemd-resolved at all and pointed /etc/resolv.conf to
    ::1 for bind9?

    Did you also check /etc/nsswitch.conf, so hosts only includes files and dns?

    There is a library for systemd-resolved that is called first.

    --
    Gruá
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Wednesday, July 15, 2026 12:30:01
    Hi,

    On Wed, Jul 15, 2026 at 11:28:56AM +0200, Marco Moock wrote:
    It is a bug in systemd-resolve, I noticed that and blamed sendmail first...

    Can it still be this even though OP reproduced it by directly querying
    their bind9 server with dig?

    There are so many bugs in systemd-resolved though, it's quite
    disappointing.

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Chris Green@3:633/10 to All on Wednesday, July 15, 2026 13:20:01
    Andy Smith <andy@strugglers.net> wrote:
    Hi,

    On Wed, Jul 15, 2026 at 11:28:56AM +0200, Marco Moock wrote:
    It is a bug in systemd-resolve, I noticed that and blamed sendmail first...

    Can it still be this even though OP reproduced it by directly querying
    their bind9 server with dig?

    There are so many bugs in systemd-resolved though, it's quite
    disappointing.

    Not just bugs, mis-features as well. I used to disable systemd-resolved
    and use dnsmasq instead on my systems but now all my DNS is done by my
    Asus router using dnsmasq on that. All I have on my Debian (and
    Raspbian) systems is local DNS caching.

    --
    Chris Green
    ú

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Wednesday, July 15, 2026 13:20:02
    On Wed, 15 Jul 2026 07:08:04 +0000
    Andy Smith <andy@strugglers.net> wrote:



    What does:

    dig +trace -x 82.195.75.100 @192.168.71.8

    output?

    It will show you each step of the recursive resolving.

    Do note that part of this recursion uses DNSSEC. If validation fails
    for some reason then SERVFAIL is the usual outcome. BIND9 as a
    resolver does do DNSSEC validation by default, whereas your ISP
    router's resolver probably does not. So your DNSSEC validation setup
    may be iffy.

    joe@jrenewsid:~$ sudo dig +trace -x 82.195.75.100 @192.168.71.8

    ; <<>> DiG 9.20.24-1+b1-Debian <<>> +trace -x 82.195.75.100
    @192.168.71.8 ;; global options: +cmd
    468193 IN NS l.root-servers.net. .
    468193 IN NS d.root-servers.net. .
    468193 IN NS f.root-servers.net. .
    468193 IN NS h.root-servers.net. .
    468193 IN NS j.root-servers.net. .
    468193 IN NS b.root-servers.net.
    468193 IN NS a.root-servers.net. .
    468193 IN NS e.root-servers.net. .
    468193 IN NS m.root-servers.net. .
    468193 IN NS g.root-servers.net. .
    468193 IN NS i.root-servers.net. .
    468193 IN NS c.root-servers.net.
    .468193 IN NSk.root-servers.net. .
    468193 IN RRSIG NS 8 0 518400 20260727190000
    20260714180000 57780 .
    1qLr5dmWdIvW7LA3BF+R/+O4gdMyZot2CdXC6TFDc9mq+pwbXvnFBeg X uR1pqR2ZUvWQm6Sgnbyra0BobPh+uAY0fZVfX6HvNZxeLDIkRNuz0qow 0NTBiu7dM122k+l66s6u6mIn0ApAmUloJs4k1b5qhNXnfLCufvcxmSJk LmebGMJAumT1F608zvuxOaZ7tcTE5Puios6H/fB rR2aXWxXiewYqfj98 gzeZ42jKjYYJMiz8P6QK5iOFN37a4ARqFFiIG8OPXPQJ31jOTN9v38h8 ApvHVEpexuV/WsVbKWiXGoxUmxynv1/bSV71UnbBEEeRDFPgkaKVSXul /ZouAw==
    ;; Received 1137 bytes from 192.168.71.8#53(192.168.71.8) in 4 ms

    ;; UDP setup with 2001:500:a8::e#53(2001:500:a8::e) for 100.75.195.82.in-addr.arpa. failed: network unreachable.
    ;; no servers could be reached
    ;; UDP setup with 2001:500:a8::e#53(2001:500:a8::e) for 100.75.195.82.in-addr.arpa. failed: network unreachable.
    ;; no servers could be reached
    ;; UDP setup with 2001:500:a8::e#53(2001:500:a8::e) for 100.75.195.82.in-addr.arpa. failed: network unreachable.
    ;; UDP setup with 2001:500:a8::e#53(2001:500:a8::e) for 100.75.195.82.in-addr.arpa. failed: network unreachable.
    in-addr.arpa. 172800 IN NS a.in-addr-servers.arpa.
    in-addr.arpa. 172800 IN NS e.in-addr-servers.arpa.
    in-addr.arpa. 172800 IN NS c.in-addr-servers.arpa.
    in-addr.arpa. 172800 IN NS d.in-addr-servers.arpa.
    in-addr.arpa. 172800 IN NS b.in-addr-servers.arpa.
    in-addr.arpa. 172800 IN NS f.in-addr-servers.arpa.
    in-addr.arpa. 86400 IN DS 47054 8 2 5CAFCCEC201D1933B4C9F6A9C8F51E51F3B39979058AC21B8DF1B1F2 81CBC6F2
    in-addr.arpa. 86400 IN DS 54956 8 2 E0E2BF5CFBD66572CA05EC18267D91509BA6A9405AF05C3FD4141DFA 45200C08
    in-addr.arpa. 86400 IN DS 53696 8 2 13E5501C56B20394DA921B51412D48B7089C5EB6957A7C58553C4D4D 424F04DF
    in-addr.arpa. 86400 IN DS 63982 8 2 AAF4FB5D213EF25AE44679032EBE3514C487D7ABD99D7F5FEC3383D0 30733C73
    in-addr.arpa. 86400 IN RRSIG DS 8 2
    86400 20260727180000 20260714170000 39561 arpa. ggoV/lumzdcvYdHm0Pm1nFSsHMBUEIYLd1kZHfIk9cD9sRadr9na iJb0 6nuPbt2UaW64gZvBs1OjCGmSM/gwVTdJj3gtOMdyXaxwb0ebYwhrrzBf ckDVLwkZd3wKfs/Ym6gb1zrrNPCNyNQKpXkhsLnukCx4TG2YIBxuH0KR 1cg7fyrZ63kLKmIks4dCmNUUV76u9OnrCQcW 05qSvsgopIOfoHjhlq1H KxfpdVPJwPMGnFBIHhtYGEimYt4W5CKBQ5RCd6rEJQ5IE+Vn0mH8oWbL E9SdMJ5rbN7144vh0rHeHdGhu9tM6FmOvsg15k5q6BiGyQy29RagK7OX BVFsmA== ;;
    Received 943 bytes from 192.33.4.12#53(c.root-servers.net) in 16 ms

    ;; UDP setup with 2001:43f8:110::10#53(2001:43f8:110::10) for 100.75.195.82.in-addr.arpa. failed: network unreachable.
    82.in-addr.arpa. 86400 IN NS ns3.lacnic.net.
    82.in-addr.arpa. 86400 IN NS ns3.afrinic.net.
    82.in-addr.arpa. 86400 IN NS ns4.apnic.net.
    82.in-addr.arpa. 86400 IN NS pri.authdns.ripe.net.
    82.in-addr.arpa. 86400 IN NS rirns.arin.net.
    82.in-addr.arpa. 86400 IN DS 64189 13 2 C1D0F2EB646A851553495E12567C265C5DB9811C61C822DA5C5BD6A6 7F38C2B4 82.in-addr.arpa. 86400 IN RRSIG DS 8 3
    86400 20260801224726 20260712091556 31765 in-addr.arpa. IfRj517uDX81sM2h0alT+sxvuypq7XJVMmTbAoVpnUEc CJ1h1LLLluyJ EWm4NxMn3G4UyQ24/Pn8AzD4FQcUJD/Z8RuI1DavhcvVxTZehvIxhcSW 3AUjmiVuT4xMdzfL9hcnN6d9ntuKjJLRsGnl6n3hAK7y/tcRsYIG8TFN hdI= ;;
    Received 409 bytes from 193.0.9.1#53(f.in-addr-servers.arpa) in 8 ms

    75.195.82.in-addr.arpa. 86400 IN NS ns2.man-da.de.
    75.195.82.in-addr.arpa. 86400 IN NS ns3.man-da.de.
    75.195.82.in-addr.arpa. 86400 IN NS ns1.man-da.de.
    75.195.82.in-addr.arpa. 3600 IN DS 57319 13 2 03644989133634C2E95982379AA9C243FE48125F67AC7039BF0E4B1B 82477E22 75.195.82.in-addr.arpa. 3600 IN RRSIG DS 13
    5 3600 20260727095121 20260713082121 64189 82.in-addr.arpa. ppXKEizvrNkimlP6xhfuVBNgGVLQ2UqBPNHe7NiTL DI2849NOsXnehYo zUn34MRuGqzp1WZoaHzDIuxlCFQCtA==
    ;; Received 305 bytes from 202.12.31.53#53(ns4.apnic.net) in 216 ms

    ;; Connection to 2001:4c80:0:7f1::249#53(2001:4c80:0:7f1::249) for 100.75.195.82.in-addr.arpa. failed: network unreachable.
    ;; Connection to 2001:41b8:0:18::10#53(2001:41b8:0:18::10) for 100.75.195.82.in-addr.arpa. failed: network unreachable. 100.75.195.82.in-addr.arpa. 3600 IN CNAME
    100.64-26.75.195.82.in-addr.arpa.
    100.75.195.82.in-addr.arpa. 3600
    IN RRSIG CNAME 13 6 3600 20260725072209 20260711055209
    25684 75.195.82.in-addr.arpa. e7FvHXk4Bqz9JLxenWVC4CDMZCj5wsD NfqrAD1Au67FcttQtcJJQGwZi YnyXONXn2HfYMcAvw8wOT4VjMWi0Fw== 64-26.75.195.82.in-addr.arpa. 3600 IN NS rns1.debian.org. 64-26.75.195.82.in-addr.arpa. 3600 IN NS rns2.debian.com. 64-26.75.195.82.in-addr.arpa. 3600 IN NS rns3.debian.org. 64-26.75.195.82.in-addr.arpa. 3600 IN NS rns4.debian.com. lp0k9coo0o6pt6r38q4jtplcdsuf9tai.75.195.82.in-addr.arpa. 86400 IN
    NSEC3 1 0 0 - LU919LMT93UQJRQO6A2ETA4E84SMOAQ0 NS lp0k9coo0o6pt6r38q4jtplcdsuf9tai.75.195.82.in-addr.arpa. 86400 IN
    RRSIG NSEC3 13 6 86400 20260725072209 20260711055209 25684 75.195.82.in-addr.arpa. yFEvhq OdRcFNZ8eDHY0MQ1DOYEOoNqqJ0z1wQUzRYjB1f2dOrw4mhNQl fgtWoKmxj0fpd/ORnKANNEDllwux7g== ;; Received 511 bytes from 82.195.66.249#53(ns1.man-da.de) in 28 ms



    Success despite some problems on the way.

    Anything in bind9's logs?

    Nothing, with trace used. Without it:

    Jul 15 11:35:09 piserve1 named[3722]: network unreachable resolving 'pri.authdns.ripe.net/A/IN': 2001:dd8:12::53#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving 'pri.authdns.ripe.net/A/IN': 2620:38:2000::53#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving 'pri.authdns.ripe.net/AAAA/IN': 2620:38:2000::53#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving
    'ns3.lacnic.net/A/IN': 2001:13c7:7002:3000::10#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving
    'ns3.lacnic.net/AAAA/IN': 2001:13c7:7002:3000::10#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving '195.82.in-addr.arpa/NS/IN': 2620:38:2000::53#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving
    'ns3.lacnic.net/A/IN': 2001:dd8:12::53#53 Jul 15 11:35:09 piserve1
    named[3722]: network unreachable resolving 'ns3.lacnic.net/AAAA/IN': 2620:38:2000::53#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving 'a.lactld.org/A/IN': 2801:1b8:10::b#53 Jul 15
    11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/AAAA/IN': 2801:1b8:10::b#53 Jul 15 11:35:09 piserve1
    named[3722]: network unreachable resolving 'ns3.lacnic.net/A/IN': 2620:38:2000::53#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving 'ns3.lacnic.net/AAAA/IN':
    2001:13c7:7002:3000::11#53 Jul 15 11:35:09 piserve1 named[3722]:
    network unreachable resolving 'a.lactld.org/A/IN': 2001:7fe::53#53 Jul
    15 11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/AAAA/IN': 2001:7fe::53#53 Jul 15 11:35:09 piserve1
    named[3722]: network unreachable resolving 'ns3.lacnic.net/A/IN': 2001:13c7:7002:3000::11#53 Jul 15 11:35:09 piserve1 named[3722]:
    network unreachable resolving 'a.lactld.org/A/IN':
    2001:503:c27::2:30#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving 'a.lactld.org/AAAA/IN': 2001:503:c27::2:30#53 Jul
    15 11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/A/IN': 2001:500:2f::f#53 Jul 15 11:35:09 piserve1
    named[3722]: network unreachable resolving 'a.lactld.org/AAAA/IN': 2001:500:2f::f#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving 'a.lactld.org/A/IN': 2001:500:9f::42#53 Jul 15
    11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/AAAA/IN': 2001:500:9f::42#53 Jul 15 11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/A/IN':
    2001:dc3::35#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving 'a.lactld.org/AAAA/IN': 2001:dc3::35#53 Jul 15
    11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/A/IN': 2001:503:ba3e::2:30#53 Jul 15 11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/AAAA/IN': 2001:503:ba3e::2:30#53 Jul 15 11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/A/IN': 2001:7fd::1#53 Jul 15
    11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/AAAA/IN': 2001:7fd::1#53 Jul 15 11:35:09 piserve1
    named[3722]: network unreachable resolving 'a.lactld.org/A/IN': 2001:500:12::d0d#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving '75.195.82.in-addr.arpa/NS/IN':
    2a13:27c0:30::5#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving 'a.lactld.org/AAAA/IN': 2001:500:12::d0d#53 Jul
    15 11:35:09 piserve1 named[3722]: network unreachable resolving 'a.lactld.org/A/IN': 2001:500:1::53#53 Jul 15 11:35:09 piserve1
    named[3722]: network unreachable resolving 'a.lactld.org/AAAA/IN': 2001:500:1::53#53 Jul 15 11:35:09 piserve1 named[3722]: network
    unreachable resolving 'a2.org.afilias-nst.info/A/IN': 2001:7fd::1#53
    Jul 15 11:35:09 piserve1 named[3722]: network unreachable resolving 'a0.org.afilias-nst.info/A/IN': 2001:7fd::1#53 Jul 15 11:35:09 piserve1 named[3722]: network unreachable resolving
    'a2.org.afilias-nst.info/A/IN': 2001:500:9f::42#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving 'a2.org.afilias-nst.info/AAAA/IN': 2001:500:9f::42#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving 'a0.org.afilias-nst.info/A/IN': 2001:500:9f::42#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving 'a0.org.afilias-nst.info/AAAA/IN': 2001:500:9f::42#53 Jul 15 11:35:09
    piserve1 named[3722]: network unreachable resolving
    'a.lactld.org/A/IN': 2001:500:48::1#53


    Presumably the router is forwarding to the ISP's master caching DNS
    server, which isn't having problems. But it looks as if the trouble is
    outside my network, and there's nothing I can do to fix it without
    forwarding to the ISP's server myself. I don't like using router DNS,
    once I had some trouble sending email to a few people which turned out
    to be the router not finding an MX record for a few domains. Since then
    I've used root hints. I still don't understand why I had no trouble
    with (presumably an older) BIND9 on the Microserver.

    Maybe a clue: my ISP doesn't yet do IPv6, though I can't see why that
    should affect a DNS enquiry for an IPv4 address.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Wednesday, July 15, 2026 13:40:01
    Hi,

    On Wed, Jul 15, 2026 at 12:16:14PM +0100, Joe wrote:
    ;; UDP setup with 2001:500:a8::e#53(2001:500:a8::e) for 100.75.195.82.in-addr.arpa. failed: network unreachable.

    [?]

    Maybe a clue: my ISP doesn't yet do IPv6, though I can't see why that
    should affect a DNS enquiry for an IPv4 address.

    You definitely have a problem with IPv6 in that your DNS resolver
    machine is trying to communicate with IPv6 addresses while not being
    able to. This indicates that the machine thinks it has a global scope
    IPv6 address, but it has no IPv6 route. It thinks it SHOULD be able to
    use IPv6 but it will always fail.

    You should either fix IPv6 or else remove the IPv6 address - IPv6 will
    always be tried first so it's making everything slower for you. I don't
    think it is the root of your problem though. I still think there is some misconfiguration of your bind9 server, as this thing is resolvable by
    everyone else.

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marco Moock@3:633/10 to All on Wednesday, July 15, 2026 15:30:02
    Am 15.07.26 um 12:30 schrieb Andy Smith:
    Can it still be this even though OP reproduced it by directly querying
    their bind9 server with dig?

    Not in this case, but in case the libc resolver is being used, it can
    call systemd-resolve via the library, see nsswitch.conf.

    --
    Gruá
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Wednesday, July 15, 2026 16:30:01
    On Wed, 15 Jul 2026 11:30:52 +0000
    Andy Smith <andy@strugglers.net> wrote:

    Hi,

    On Wed, Jul 15, 2026 at 12:16:14PM +0100, Joe wrote:
    ;; UDP setup with 2001:500:a8::e#53(2001:500:a8::e) for 100.75.195.82.in-addr.arpa. failed: network unreachable.

    [?]

    Maybe a clue: my ISP doesn't yet do IPv6, though I can't see why
    that should affect a DNS enquiry for an IPv4 address.

    You definitely have a problem with IPv6 in that your DNS resolver
    machine is trying to communicate with IPv6 addresses while not being
    able to. This indicates that the machine thinks it has a global scope
    IPv6 address, but it has no IPv6 route. It thinks it SHOULD be able to
    use IPv6 but it will always fail.

    Not up to me, the router apparently has access to IPv6 but doesn't pass
    it through, though from the menu it looks capable of doing it. But
    there's no option to enable it, and the ISP (Plusnet, owned by BT) says
    a trial is under way.

    But there's no errors in the log with the trace option running, even
    though something further along is getting network unreachable messages,
    and the correct data is returned. Without trace the local BIND9 is
    attempting use of IPv6, which it obviously isn't with trace running.

    You should either fix IPv6 or else remove the IPv6 address - IPv6 will
    always be tried first so it's making everything slower for you. I
    don't think it is the root of your problem though. I still think
    there is some misconfiguration of your bind9 server, as this thing is resolvable by everyone else.

    That's what I hoped, but it is the Debian original setup plus local
    forward and reverse zones, as I have always done. I used to link it to
    DHCP to get dynamic updates, but I don't bother these days.

    I don't make any use of IPv6 in my network, so can I just disable it at
    the operating system level? At least until Plusnet roll it out to its customers.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Wednesday, July 15, 2026 16:50:01
    Hi,

    On Wed, Jul 15, 2026 at 03:28:51PM +0100, Joe wrote:
    I don't make any use of IPv6 in my network, so can I just disable it at
    the operating system level? At least until Plusnet roll it out to its customers.

    You can disable SLAAC in whatever sets up your networking. Or you can
    disable IPv6 entirely by booting with "ipv6.disable=1" on the kernel
    command line.

    Either way will result in you not getting a global scope IPv6 address
    which will stop applications on your server trying to use IPv6.

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marco Moock@3:633/10 to All on Thursday, July 16, 2026 09:00:01
    Am 15.07.26 um 16:30 schrieb Joe:
    I don't make any use of IPv6 in my network, so can I just disable it at
    the operating system level? At least until Plusnet roll it out to its customers.

    No, this is a bad idea. You need to figure out why your system has a non-working GUA address.

    Show the output of
    ip a
    ip -6 route show


    --
    Gruá
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Thursday, July 16, 2026 09:10:01
    On Thu, 16 Jul 2026 08:39:29 +0200
    Marco Moock <mm@dorfdsl.de> wrote:

    Am 15.07.26 um 16:30 schrieb Joe:
    I don't make any use of IPv6 in my network, so can I just disable
    it at the operating system level? At least until Plusnet roll it
    out to its customers.

    No, this is a bad idea. You need to figure out why your system has a non-working GUA address.

    Show the output of
    ip a
    ip -6 route show



    My ISP does not currently provide IPv6 connection to the Internet.

    Nothing I can do about it.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Thursday, July 16, 2026 09:20:01
    Hi,

    On Thu, Jul 16, 2026 at 08:00:52AM +0100, Joe wrote:
    My ISP does not currently provide IPv6 connection to the Internet.

    Nothing I can do about it.

    The router is misconfigured if it is handing out addresses and IPv6
    default route that do not work (or no default route). Possibly there is
    a way you can fix that, and that would be the best place to fix it. If
    you can't fix it then you will need to either ignore SLAAC or disable
    IPv6 as I mentioned in the prior reply.

    The thing is that if you have a router handing out IPv6 addresses and
    route by SLAAC then potentially every device on your network will be
    trying to use IPv6, which is obviously bad if that's not going to work. Disabling it on every host is also clearly more work.

    It is not normal for a Plusnet router to be giving out SLAAC details
    that don't work. This will cause problems. Either customer services or
    other customers (in some sort of forum) should be able to advise you how
    to disable SLAAC in the router if it's not going to work. "We don't
    support IPv6" can't be the end of the conversation, because this is
    broken.

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Thursday, July 16, 2026 10:40:01
    On Thu, 16 Jul 2026 07:10:15 +0000
    Andy Smith <andy@strugglers.net> wrote:

    Hi,

    On Thu, Jul 16, 2026 at 08:00:52AM +0100, Joe wrote:
    My ISP does not currently provide IPv6 connection to the Internet.

    Nothing I can do about it.

    The router is misconfigured if it is handing out addresses and IPv6
    default route that do not work (or no default route). Possibly there
    is a way you can fix that, and that would be the best place to fix
    it. If you can't fix it then you will need to either ignore SLAAC or
    disable IPv6 as I mentioned in the prior reply.

    The router menus make no mention of IPv6 configuration, DHCP refers
    only to IPv4 and is working. One of the Pis has Kea installed but with
    no interfaces configured for IPv4 or IPv6.

    All three of my Linux computers have IPv6 addresses, the sid
    workstation having two. I've never used it so I've no idea where they
    have come from. Neither /etc/network nor Network Manager have any IPv6 configurations.

    The thing is that if you have a router handing out IPv6 addresses and
    route by SLAAC then potentially every device on your network will be
    trying to use IPv6, which is obviously bad if that's not going to
    work. Disabling it on every host is also clearly more work.

    It is not normal for a Plusnet router to be giving out SLAAC details
    that don't work. This will cause problems. Either customer services or
    other customers (in some sort of forum) should be able to advise you
    how to disable SLAAC in the router if it's not going to work. "We
    don't support IPv6" can't be the end of the conversation, because
    this is broken.

    Thanks,
    Andy


    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marco Moock@3:633/10 to All on Thursday, July 16, 2026 11:20:02
    Am 16.07.26 um 10:40 schrieb Joe:
    All three of my Linux computers have IPv6 addresses, the sid
    workstation having two. I've never used it so I've no idea where they
    have come from. Neither /etc/network nor Network Manager have any IPv6 configurations.

    They will generate link-local addresses automatically, but that is
    intended, no issue.
    It will never use them for outgoing traffic to GUA addresses.

    Does it have other addresses assigned?


    --
    Gruá
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Thursday, July 16, 2026 13:20:02
    On Thu, 16 Jul 2026 10:54:53 +0200
    Marco Moock <mm@dorfdsl.de> wrote:

    Am 16.07.26 um 10:40 schrieb Joe:
    All three of my Linux computers have IPv6 addresses, the sid
    workstation having two. I've never used it so I've no idea where
    they have come from. Neither /etc/network nor Network Manager have
    any IPv6 configurations.

    They will generate link-local addresses automatically, but that is
    intended, no issue.
    It will never use them for outgoing traffic to GUA addresses.

    Does it have other addresses assigned?



    All three have IPv4 addresses, one assigned by the router and the
    others static. No problems using IPv4. The dig command using trace does
    not attempt to use IPv6 out of the network, but presumably without
    trace, one of the intermediate DNS servers actually requires IPv6, and
    my server tries to use it.

    Actually all three have two IPv6 addresses. Presumably they will be
    used within the network if necessary.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Jeffrey Walton@3:633/10 to All on Thursday, July 16, 2026 14:40:01
    On Wed, Jul 15, 2026 at 10:29?AM Joe <joe@jretrading.com> wrote:

    On Wed, 15 Jul 2026 11:30:52 +0000
    Andy Smith <andy@strugglers.net> wrote:

    Hi,

    On Wed, Jul 15, 2026 at 12:16:14PM +0100, Joe wrote:
    ;; UDP setup with 2001:500:a8::e#53(2001:500:a8::e) for 100.75.195.82.in-addr.arpa. failed: network unreachable.

    [?]

    Maybe a clue: my ISP doesn't yet do IPv6, though I can't see why
    that should affect a DNS enquiry for an IPv4 address.

    You definitely have a problem with IPv6 in that your DNS resolver
    machine is trying to communicate with IPv6 addresses while not being
    able to. This indicates that the machine thinks it has a global scope
    IPv6 address, but it has no IPv6 route. It thinks it SHOULD be able to
    use IPv6 but it will always fail.

    Not up to me, the router apparently has access to IPv6 but doesn't pass
    it through, though from the menu it looks capable of doing it. But
    there's no option to enable it, and the ISP (Plusnet, owned by BT) says
    a trial is under way.

    But there's no errors in the log with the trace option running, even
    though something further along is getting network unreachable messages,
    and the correct data is returned. Without trace the local BIND9 is
    attempting use of IPv6, which it obviously isn't with trace running.

    You should either fix IPv6 or else remove the IPv6 address - IPv6 will always be tried first so it's making everything slower for you. I
    don't think it is the root of your problem though. I still think
    there is some misconfiguration of your bind9 server, as this thing is resolvable by everyone else.

    That's what I hoped, but it is the Debian original setup plus local
    forward and reverse zones, as I have always done. I used to link it to
    DHCP to get dynamic updates, but I don't bother these days.

    I don't make any use of IPv6 in my network, so can I just disable it at
    the operating system level? At least until Plusnet roll it out to its customers.

    Maybe add a firewall rule at the router that blocks or drops DHCPv6.
    If a host cannot obtain an IPv6 address, it should effectively stop
    most of the traffic causing the problems. It looks like UDP/546 and
    UDP/547 are the ports of interest.

    Or you can consider standing up a pfSense firewall between your router
    and your network. You will have a lot more control over what happens
    on your network. You will have complete admin access, so you can
    disable services like DHCPv6. And the pfSense box can perform the
    necessary translations to tunnel IPv6 through IPv4, if you really want
    IPv6 on your internal network.

    Jeff

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Thursday, July 16, 2026 20:10:01
    On Thu, 16 Jul 2026 13:51:18 -0400
    Lee <ler762@gmail.com> wrote:

    On Thu, Jul 16, 2026 at 3:01?AM Joe wrote:

    On Thu, 16 Jul 2026 08:39:29 +0200
    Marco Moock wrote:

    Am 15.07.26 um 16:30 schrieb Joe:
    I don't make any use of IPv6 in my network, so can I just
    disable it at the operating system level? At least until
    Plusnet roll it out to its customers.

    No, this is a bad idea. You need to figure out why your system
    has a non-working GUA address.

    Show the output of
    ip a
    ip -6 route show



    My ISP does not currently provide IPv6 connection to the Internet.

    And the problem is that named is trying to use IPv6 to query name
    servers on the Internet and that fails? Correct?

    you can tell named to use only IPv4 by editing
    /etc/default/named
    and adding "-4" to the OPTIONS variable - eg:

    # startup options for the server
    OPTIONS="-4 -u bind"


    Thank you, I was going to look in that direction in a while. Currently
    copying about 60GB from backup to one of the Pis and trying to get rid
    of googleusercontent from my Inbox.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Friday, July 17, 2026 13:10:01
    Hi,

    On Thu, Jul 16, 2026 at 07:03:51PM +0100, Joe wrote:
    On Thu, 16 Jul 2026 13:51:18 -0400
    Lee <ler762@gmail.com> wrote:
    you can tell named to use only IPv4 by editing
    /etc/default/named
    and adding "-4" to the OPTIONS variable - eg:

    # startup options for the server
    OPTIONS="-4 -u bind"

    Thank you, I was going to look in that direction in a while

    Disabling IPv6 in one app on one machine seems like a terrible idea when
    the problem is that you have broken IPv6 on every machine in your
    network.

    If you can't fix the router why would you not just disable SLAAC or IPv6
    on each machine?

    A lot of work seems to be being made here out of a simple problem.

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Friday, July 17, 2026 15:10:01
    On Fri, 17 Jul 2026 11:00:22 +0000
    Andy Smith <andy@strugglers.net> wrote:

    Hi,

    On Thu, Jul 16, 2026 at 07:03:51PM +0100, Joe wrote:
    On Thu, 16 Jul 2026 13:51:18 -0400
    Lee <ler762@gmail.com> wrote:
    you can tell named to use only IPv4 by editing
    /etc/default/named
    and adding "-4" to the OPTIONS variable - eg:

    # startup options for the server
    OPTIONS="-4 -u bind"

    Thank you, I was going to look in that direction in a while

    Disabling IPv6 in one app on one machine seems like a terrible idea
    when the problem is that you have broken IPv6 on every machine in your network.

    If you can't fix the router why would you not just disable SLAAC or
    IPv6 on each machine?

    A lot of work seems to be being made here out of a simple problem.


    Can I ask what the advantages are to using IPv6 inside a four-computer
    network?

    I thought I had said that there is probably nothing wrong with the
    router, I just don't have an IPv6 connection to the Internet, because
    my ISP does not currently provide it. When that changes, the inability
    of BIND9 to see the Debian server will disappear. Until then, I will
    grudgingly use my ISP's DNS servers.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Friday, July 17, 2026 15:50:02
    On Fri, Jul 17, 2026 at 02:06:51PM +0100, Joe wrote:
    On Fri, 17 Jul 2026 11:00:22 +0000
    Andy Smith <andy@strugglers.net> wrote:
    Disabling IPv6 in one app on one machine seems like a terrible idea
    when the problem is that you have broken IPv6 on every machine in your network.

    If you can't fix the router why would you not just disable SLAAC or
    IPv6 on each machine?

    A lot of work seems to be being made here out of a simple problem.

    Can I ask what the advantages are to using IPv6 inside a four-computer network?

    I am baffled how you do not seem to grasp that what you have right now
    is BROKEN AND CAUSING PROBLEMS. Your computers think they have IPv6 and
    so are trying to use it. Linux chooses to use IPv6 first if it thinks it
    can. Since yours doesn't work, every host you connect to that has an
    IPv6 address in DNS introduces a delay while your computers figure out
    that it isn't working and retry on IPv4.

    I have suggested at least three times now that you disable IPv6 if you
    can't make it work. That would be better than having it in a broken
    state.

    It's besides the point what the advantages might be of using IPv6 on
    your LAN. You don't have working IPv6. What you have is a broken setup
    that is causing you problems every time you resolve a host that has IPv6 records in the DNS. Fix that first and think about IPv6 on the LAN
    later, if ever.

    I thought I had said that there is probably nothing wrong with the
    router

    There 100% is a problem with your router. It is most likely sending IPv6 autoconfig packets that cause your hosts to autoconfigure an IPv6
    address by SLAAC. If it's not doing that then it is doing DHCPv6, but I
    think SLAAC is more likely.

    There is no other way for your hosts to have global scope IPv6 addresses
    unless you statically configured them, which you say you did not. We
    could tell more if you would simply show us the full output of "ip addr
    show", which you have also been asked for several times now.

    I just don't have an IPv6 connection to the Internet, because
    my ISP does not currently provide it.

    If your connection does not support IPv6 then it is NOT CORRECT for your
    router to be handing out IPv6 addresses. It is NOT EXPECTED for your
    hosts to have IPv6 addresses in this situation. The fact that they do is causing you problems. I don't know how to make this more clear.

    When that changes, the inability of BIND9 to see the Debian server
    will disappear.

    That has actually not been proven yet, but at this point is a bit of
    a tangent.

    Until then, I will grudgingly use my ISP's DNS servers.

    That is not the issue. Any DNS resolver, including third party ones run
    by Google or whoever, will hand you back AAAA records (IPv6 addresses),
    which your hosts will try to reach if they think they have IPv6
    connectivity. This happens even if you yourself do not have any IPv6 connectivity. The difference is that your apps do not try to use any
    AAAA record if they themselves don't have IPv6. Yours think they do, and
    that's a problem.

    You must stop your hosts from having global scope IPv6 addresses if IPv6
    is not going to work. They WILL try to use them otherwise.

    BTW, making bind9 use only IPv6 will not stop it handing back AAAA
    records in answers, so other software on your network will still
    encounter problems if it's using your bind9 server as a resolver.

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marco Moock@3:633/10 to All on Friday, July 17, 2026 18:20:01
    Am 16.07.26 um 13:20 schrieb Joe:
    Actually all three have two IPv6 addresses. Presumably they will be
    used within the network if necessary.

    Please run ip a and show the output here.

    There are different types and if it tries to use link-local or ULA for accessing GUA addresses, it will not work.

    --
    Gruá
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marco Moock@3:633/10 to All on Friday, July 17, 2026 18:20:01
    Am 17.07.26 um 15:10 schrieb Joe:
    Can I ask what the advantages are to using IPv6 inside a four-computer network?

    I thought I had said that there is probably nothing wrong with the
    router, I just don't have an IPv6 connection to the Internet, because
    my ISP does not currently provide it.

    In such a case a machine only has link-local addresses. They will never
    be used to connect to GUA (public) addresses.
    If there are other addresses attached, this is an issue that needs to be examined.

    --
    Gruá
    Marco

    Junk-Mail bitte an trashcan@stinkedores.dorfdsl.de

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Greg Wooledge@3:633/10 to All on Friday, July 17, 2026 19:40:01
    On Fri, Jul 17, 2026 at 17:57:27 +0200, Marco Moock wrote:
    Am 17.07.26 um 15:10 schrieb Joe:
    I thought I had said that there is probably nothing wrong with the
    router, I just don't have an IPv6 connection to the Internet, because
    my ISP does not currently provide it.

    In such a case a machine only has link-local addresses. They will never be used to connect to GUA (public) addresses.
    If there are other addresses attached, this is an issue that needs to be examined.

    There must be more to it than this, because my machine has a "scope
    global" IPv6 address, but no default IPv6 route, and everything is
    behaving reasonably (IPv6 does not work, as expected, but nothing tries
    to use it either).

    hobbit:~$ ip a
    ...
    2: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether d0:ad:08:23:8b:79 brd ff:ff:ff:ff:ff:ff
    altname enp0s31f6
    altname enxd0ad08238b79
    inet 192.168.1.105/24 brd 192.168.1.255 scope global dynamic noprefixroute eno1
    valid_lft 29852sec preferred_lft 24452sec
    inet6 fd39:8f22:f952:cb29:7bef:e9bc:8ae5:2f96/64 scope global dynamic mngtmpaddr noprefixroute
    valid_lft 1717sec preferred_lft 1717sec
    inet6 fe80::95a9:69c4:b41f:5f90/64 scope link
    valid_lft forever preferred_lft forever
    ...

    hobbit:~$ ip -6 route show
    fd39:8f22:f952:cb29::/64 dev eno1 proto ra metric 1002 pref medium fd78:f817:26c2:1::/64 via fe80::b7f9:2a9b:9bd3:6293 dev eno1 metric 1002 pref medium
    fe80::/64 dev eno1 proto kernel metric 256 pref medium


    My ISP is Brightspeed, and my router (provided by Brightspeed) is
    running "EXOS Version 26.1.0" according to a tooltip on the router's administrative web page. The router very clearly says that IPv6 is "Disconnected" on the first page that I see when I login.

    I have done nothing whatsoever to change how IPv6 works. I wouldn't even
    know how. This is all out-of-the-box behavior.

    For a very large number of people, IPv6 connectivity is still unavailable.

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Joe@3:633/10 to All on Friday, July 17, 2026 21:30:01
    On Fri, 17 Jul 2026 13:42:17 +0000
    Andy Smith <andy@strugglers.net> wrote:



    There 100% is a problem with your router. It is most likely sending
    IPv6 autoconfig packets that cause your hosts to autoconfigure an IPv6 address by SLAAC. If it's not doing that then it is doing DHCPv6, but
    I think SLAAC is more likely.

    Yes, you are right, I tracked it down and it was set to 'stateless',
    even though the broadband status showed 'IPv6 network' as 'disabled',
    without any means of enabling it.

    I've turned it off, rebooted the router although it didn't ask for it,
    and rebooted my workstation.

    There is no other way for your hosts to have global scope IPv6
    addresses unless you statically configured them, which you say you
    did not. We could tell more if you would simply show us the full
    output of "ip addr show", which you have also been asked for several
    times now.

    1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
    group default qlen 1000 link/loopback 00:00:00:00:00:00 brd
    00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo
    valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host noprefixroute
    valid_lft forever preferred_lft forever
    2: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel
    state UP group default qlen 1000 link/ether 74:46:a0:b2:40:1f brd ff:ff:ff:ff:ff:ff altname enp0s25
    altname enx7446a0b2401f
    inet 192.168.71.106/24 brd 192.168.71.255 scope global dynamic noprefixroute eno1 valid_lft 86338sec preferred_lft 86338sec
    inet6 fdda:67d:d74e:4ed1:3ee3:5e26:8a2b:7db4/64 scope global
    dynamic noprefixroute valid_lft 1752sec preferred_lft 1752sec
    inet6 fe80::edc9:2f64:d118:e10f/64 scope link noprefixroute
    valid_lft forever preferred_lft forever


    IPv6 addresses still there. It is certainly Debian putting them there,
    not me.

    --
    Joe

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Friday, July 17, 2026 22:30:01
    Hi,

    On Fri, Jul 17, 2026 at 08:27:05PM +0100, Joe wrote:
    1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
    group default qlen 1000 link/loopback 00:00:00:00:00:00 brd
    00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo
    valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host noprefixroute
    valid_lft forever preferred_lft forever
    2: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel
    state UP group default qlen 1000 link/ether 74:46:a0:b2:40:1f brd ff:ff:ff:ff:ff:ff altname enp0s25
    altname enx7446a0b2401f
    inet 192.168.71.106/24 brd 192.168.71.255 scope global dynamic noprefixroute eno1 valid_lft 86338sec preferred_lft 86338sec
    inet6 fdda:67d:d74e:4ed1:3ee3:5e26:8a2b:7db4/64 scope global
    dynamic noprefixroute valid_lft 1752sec preferred_lft 1752sec
    inet6 fe80::edc9:2f64:d118:e10f/64 scope link noprefixroute
    valid_lft forever preferred_lft forever

    IPv6 addresses still there. It is certainly Debian putting them there,
    not me.

    The above is a normal thing to see. The only global scope IPv6 address
    there is fdda:67d:d74e:4ed1:3ee3:5e26:8a2b:7db4 which is only a valid
    source address when talking to something else inside fc00::/7, which is
    nothing on the internet. So the network unreachable stuff you were
    seeing before should have gone away now.

    It's not a necessary thing to do, but the only way to completely remove
    all kinds of IPv6 addresses is to completely disable IPv6, which
    requires a reboot with "ipv6.disable=1" on the kernel command line.

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andy Smith@3:633/10 to All on Friday, July 17, 2026 22:30:02
    Hi,

    On Fri, Jul 17, 2026 at 01:38:08PM -0400, Greg Wooledge wrote:
    inet6 fd39:8f22:f952:cb29:7bef:e9bc:8ae5:2f96/64 scope global dynamic mngtmpaddr noprefixroute

    This is a Unique Local Address, like RFC1918 but for IPv6. These are
    within fc00::/7. They are not valid source addresses except if the
    destination is also within fc00::/7. Thus a system won't try to use them
    for external communications.

    https://en.wikipedia.org/wiki/Unique_local_address

    Thanks,
    Andy

    --
    https://bitfolk.com/ -- No-nonsense VPS hosting

    --- PyGate Linux v1.5.18
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)