Change Nameservers with Zero Downtime: Method and Checklist

A nameserver change causes no outage by itself: during propagation, the old and the new DNS answer side by side. The only thing that breaks sites is an incomplete destination zone. So the whole method fits in one sentence: make both zones identical before you switch.

Why it doesn't break (when done right)

Resolvers worldwide cache your old nameservers for the TTL's duration (often 1–48 h at the registry level). Inside that window, some traffic still asks the old zone, the rest asks the new one. If both serve identical records, nobody notices anything — the switch is atomic from the user's point of view.

The pre-switch checklist

1) Copy the zone completely: A, AAAA, CNAME, and above all the email records — MX, SPF (TXT), DKIM selectors (*._domainkey), DMARC. A forgotten DKIM is the most common failure: mail still sends but lands in spam.
2) Check the exotic records: SRV (VoIP, games), CAA (certificate issuance), verification TXTs (Google, Microsoft 365).
3) Compare what public DNS actually answers against the new zone — the "preflight": query the apex, www, and MX, and confirm identical answers.
4) Switch, then leave the old zone untouched: it's your safety net during propagation.

The one-minute rollback

If anything looks wrong, putting the old nameservers back is enough — the old zone never changed. That's why you never delete the origin zone for several days.

Frequently asked questions