About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
A record that exists in your DNS console and nowhere else usually means the parent zone delegates to a different set of nameservers. Resolvers follow the parent's NS records, not the console you happen to have open. RFC 1034 says the delegation NS records and glue "should be added to the parent zone", and "The administrators of both zones should insure that the NS and glue RRs which mark both sides of the cut are consistent and remain so."
Until those two lists name the same servers, edits on the unused side never become the answer the internet gets.
What the symptom means
The parent of example.com is the zone above it, such as com. RFC 1034 section 4.2.1 says the NS records that mark a cut at the bottom of a zone are not part of that zone's authoritative data, and they "should be exactly the same as the corresponding RRs in the top node of the subzone". The dig manual says +trace follows referrals from the root servers, and a query you aim yourself with @server goes to the server you name.
So a console can be honest and still be irrelevant. It is showing a zone that no referral reaches. The signed-zone case is different: the parent can name the right servers and still fail validation because a DS record points at an old key. That fault is "DNSSEC SERVFAIL after changing DNS provider: the stale DS record and the safe order to move a signed zone".
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| The registrar still lists the previous DNS host | Your console shows the new record; public lookups do not | dig +trace ends at different servers from the console |
| Two accounts both claim the domain, and you edited the idle one | The host you edited answers only when you query it with @ |
dig @ that host, then compare with the trace |
| NS hostnames match, but one hostname is a different machine from the one you think | The name resolves to an unexpected address | Compare the address in the trace's glue with the host you edit |
| The parent has the right servers, and a resolver cache is old | dig +trace shows your record, a normal lookup does not |
Compare the TTL saved from the parent query in step 2 |
Prove the cause before changing anything
All of these are read-only.
1. Follow the delegation from the root.
dig +trace example.com NS
The dig manual says +trace "toggles tracing of the delegation path from the root name servers" and, when it is enabled, "dig makes iterative queries to resolve the name being looked up. It follows referrals from the root servers, showing the answer from each server that was used to resolve the lookup." If you also pass @server, the manual says that argument "affects only the initial query for the root zone name servers", so leave it off for this test. +dnssec is set as well when +trace is set.
2. Save the parent's answer. In the trace, the reply above your zone is the parent referral. Query that parent server directly and keep the TTL. The manual says +ttlid displays the TTL when printing the record.
dig @ns.parent.example example.com NS +norecurse +ttlid > /root/parent-ns-example.com-2026-11-22.txt
Replace ns.parent.example with the parent server the trace showed. +norecurse clears the recursion-desired bit. Then save the registrar's own list: open its nameserver screen and copy every hostname into a second file, for example with nano /root/registrar-ns-example.com-2026-11-22.txt.
3. Ask the zone you edited.
dig @ns1.your-dns.example example.com SOA
dig @ns1.your-dns.example example.com A
If this server has the new record and the trace never lands on it, the parent is not delegating to the zone you edit.
Fix
You either edit the zone the parent already names, or you change the parent's NS set at the registrar to the servers that hold the zone you edit. Changing the registrar is the step below. For a signed domain, follow "DNSSEC SERVFAIL after changing DNS provider: the stale DS record and the safe order to move a signed zone" before you change delegation.
Before you change anything at the registrar, open /root/registrar-ns-example.com-2026-11-22.txt and check that it holds every nameserver hostname the registrar shows, together with any glue address shown beside a hostname. Add what is missing, because that file is the only saved copy.
Risk: Replacing the nameserver hostnames at the registrar moves the whole domain, including mail and the website, to the servers you list. If those servers do not hold this zone, names stop resolving. Glue that points at the wrong address sends resolvers to the wrong machine.
At the registrar, set the nameserver hostnames to the servers you queried in step 3, the ones that already answer with the zone you intend. Match the hostnames character for character, including the trailing domain. RFC 2181 section 10.3 says "The domain name used as the value of a NS resource record, or part of the value of a MX resource record must not be an alias", so a nameserver hostname must not be a CNAME. Change all the nameserver fields in one save.
To undo it: put the nameserver hostnames listed in /root/registrar-ns-example.com-2026-11-22.txt back into the registrar's nameserver fields. Resolvers can keep the set you just published until the TTL on those parent NS records expires.
Example (illustrative): the console on ns1.newdns.example shows www.example.com A 192.0.2.10. dig +trace www.example.com ends at ns1.olddns.example, and dig @ns1.olddns.example www.example.com A shows 192.0.2.20. The registrar's nameserver fields still list ns1.olddns.example and ns2.olddns.example. After they are changed to ns1.newdns.example and ns2.newdns.example, a later trace ends at the new servers and shows 192.0.2.10.
Confirm the fix
dig +trace example.com NSends at the servers you edited.dig @each of those servers returns the record you added.- A query with no
@servereventually returns the same record. Until parent NS TTLs expire, some resolvers can still show the old servers.
What to watch afterwards
- After a DNS-host move, compare the registrar's nameserver list with
dig +trace. - After you have changed the registrar, a parent referral that still shows the old servers may be a delay.
Read next
Related articles you may want next.
- TroubleshootingDNSSEC SERVFAIL after changing DNS provider: the stale DS record and the safe order to move a signed zoneDomain gives SERVFAIL after changing DNS provider? Prove a stale DS record with dig +dnssec and +cd, then fix it at the registrar in the safe order.Read article
- TroubleshootingGlue records missing: the nameserver hostname will not resolveA nameserver hostname will not resolve when glue is missing. Look up the NS set, then the address of each nameserver name, and add glue only when that name is inside the zone.Read article
- TroubleshootingDNS Changed but Users Still Reach the Old ApplicationDistinguish authoritative DNS, resolver caches and existing connections so an application migration has a measurable cutover and return plan.Read article
Risks and limits
ns.parent.examplestands in for the parent server in your trace.- The trace shows what the parent is serving now.
Sources
- RFC 1034: Domain Names - Concepts and Facilities — IETF, November 1987, accessed 22 November 2026
- BIND 9.18.44 manual pages: dig — ISC, no date shown, accessed 22 November 2026
- RFC 2181: Clarifications to the DNS Specification — IETF, July 1997, accessed 22 November 2026
This article was originally published on .