DNS SERVFAIL: what it means and how to fix it
SERVFAIL means a DNS resolver couldn't get an answer for your domain. The usual causes, DNSSEC and broken name servers, and the dig commands that find the fault.
· 7 min read · By the Spot Downtime team
SERVFAIL is a DNS response code that means "I tried to resolve this name and failed". Unlike NXDOMAIN, which says the name doesn't exist, SERVFAIL says nothing about whether it exists. The resolver hit a problem while looking it up, and your users see the site as unreachable. If your domain returns SERVFAIL, the fault is almost always on the authoritative side: your name servers or your DNSSEC setup.
For the NXDOMAIN case, see DNS_PROBE_FINISHED_NXDOMAIN.
Confirm it is SERVFAIL
Ask a public resolver and read the status in the header:
dig example.com A @8.8.8.8
;; ->>HEADER<<- opcode: QUERY, status: SERVFAIL, id: 41822Then ask more than one resolver. Different answers are informative:
dig example.com A @8.8.8.8 # Google
dig example.com A @1.1.1.1 # Cloudflare
dig example.com A @9.9.9.9 # Quad9- All fail: your zone or delegation is broken. Keep reading.
- Some fail, some work: often DNSSEC, because resolvers differ in validation behaviour, or one of several name servers is failing and the resolvers are picking different ones.
- Only your local resolver fails: the problem is your ISP, router or corporate DNS, not your domain.
Cause 1: DNSSEC validation failure
This is the most common reason a domain that "looks fine" returns SERVFAIL. A validating resolver checks the signatures on your records. If they don't verify, it refuses to give an answer and returns SERVFAIL rather than risk serving forged data.
The telltale test is to turn validation off for one query with +cd (checking disabled):
dig example.com A @8.8.8.8 +cd
# Works with +cd but fails without it? It is DNSSEC.
# Trace the chain of trust from the root
delv example.com A @8.8.8.8If the query works with +cd, the data is there but the signatures don't validate. Typical reasons:
- A stale DS record. You moved DNS to a new provider, or rotated keys, but the DS record at the registrar still points at the old key. The parent zone says "this domain is signed with key X" and the new servers sign with key Y. Fix: either update the DS record to match the new provider's key, or remove the DS record at the registrar if you are turning DNSSEC off. Remove the DS first and wait for its TTL to pass before decommissioning the old provider.
- Expired signatures. RRSIG records have an expiry date. If your signing software or provider stopped re-signing the zone, signatures lapse and every validating resolver fails at once.
- Unsigned secondary servers. A zone signed on one name server and served unsigned from another breaks validation intermittently, depending on which server a resolver asks.
Moving DNS providers with DNSSEC on
Cause 2: lame or unreachable name servers
If the registrar lists name servers that don't answer, or answer without being authoritative for your zone (a lame delegation), the resolver exhausts its options and returns SERVFAIL. This happens after an account lapses at the old DNS host, a zone is deleted by accident, or a typo goes into the NS records.
# What does the parent zone say your name servers are?
dig example.com NS +trace
# Ask each listed name server directly
dig example.com A @ns1.dnsprovider.com +norecurse
dig example.com A @ns2.dnsprovider.com +norecurseA healthy authoritative server answers with the aa flag in the header. No answer, a timeout or REFUSED from a listed server means that server is the problem. Remove it from the delegation or restore the zone on it.
Mismatched NS records
The NS records at the registrar (the delegation) and the NS records inside your zone should agree. When they don't, some resolvers get confused. Compare dig example.com NS @a.gtld-servers.net (the parent, for .com) with dig example.com NS @ns1.dnsprovider.com (the zone itself).
Cause 3: other zone problems
- A CNAME chain that ends nowhere. A CNAME pointing to a name whose own servers are broken returns SERVFAIL for your name, even though your zone is fine. Resolve the target directly.
- Broken glue records. If your name servers are inside the domain they serve (
ns1.example.comforexample.com), the registrar must hold their IP addresses as glue. Missing or wrong glue creates a loop the resolver can't escape. - Oversized responses. Very large answers over UDP can fail when a firewall blocks DNS over TCP or fragmented packets. Allow both UDP and TCP on port 53.
- A provider outage. If dig shows your name servers timing out and your DNS host reports an incident, wait or fail over to a secondary provider.
Cause 4: your own resolver
If public resolvers work and yours doesn't, the resolver is at fault: a stopped Unbound or BIND service, a corrupt cache, or a broken upstream. Restart it, flush its cache, and check its logs. On a Linux server, resolvectl query example.com shows which resolver the system is using and what it got back.
How long until it is fixed?
Resolvers cache SERVFAIL for a short time, usually seconds to a few minutes, so a repaired zone recovers quickly. The slower part is DNSSEC: a fixed DS record has to propagate through the parent zone and resolver caches, which can take as long as the DS record's TTL. Check what resolvers around the world see rather than waiting blindly.
DNS propagation checkerSee whether resolvers in different locations return an answer or SERVFAIL for your domain.You can also inspect the records themselves with the DNS lookup tool. More background is in DNS propagation explained and DNS TTL explained.
Catch it before customers do
SERVFAIL tends to start at a moment someone changed something: a provider move, a key rollover, a deleted zone. It is also the kind of failure that hits some users and not others. A DNS monitor that checks your records on a schedule and alerts when they stop resolving or change unexpectedly tells you within minutes, instead of when customers write in.
Get new guides by email
Practical posts on monitoring, outages, error fixes and on-call, when they're published. No spam.
Keep reading
- DNS · Troubleshooting · IncidentsWhat to do when your DNS provider goes downWhen your DNS host has an outage, your site disappears even though your servers are fine. How to confirm it, what you can do during it, and how to prepare.November 12, 2026 · 6 min read
- DNS · TroubleshootingCan you use a CNAME on a root domain? Apex records, ALIAS and flatteningA CNAME can't share a name with other records, so it breaks at the root domain. Why, what to use instead (A, ALIAS, ANAME, CNAME flattening), and how to point example.com at a host.November 2, 2026 · 6 min read