What is DNS propagation and why does it take so long?

After registering a domain and uploading your website, you may still see the old site or get a lookup error. DNS publication and caching can explain why different visitors see a change at different times.

Your DNS provider publishes the records for your domain on authoritative nameservers. Your registrar updates the parent-zone delegation to identify those nameservers. These can be separate companies from your web host.

What is DNS?

DNS stands for Domain Name System. It translates a name entered in a browser into an address the browser can use to reach a website. It also provides records for mail and other services. See DNS basics for record types and examples.

When you change a record, your provider first publishes it to its authoritative nameservers. Recursive resolvers used by visitors may still have a cached copy of the previous answer. Your provider's publication process and those caches together determine when the change becomes visible.

Factors affecting DNS delay

Your TTL (Time to Live) settings — TTL is the normal cache lifetime for an answer. With a one-hour TTL, a resolver that fetched the old answer just before a change can normally retain it for nearly another hour. Lowering the TTL helps newly fetched answers expire sooner; allow the previous TTL to elapse before making a planned change. Shorter lifetimes increase DNS lookup traffic and load on authoritative nameservers. An overloaded nameserver can respond more slowly.

Your ISP or recursive resolver — Resolvers cache answers to reduce latency and DNS traffic. Each cache expires independently, so two visitors can see different answers during a change. A previously missing name or record can also be cached as a negative answer. Some resolvers serve stale answers during authoritative-server failures to preserve availability.

Your domain registrar and registry — Changing nameservers updates the parent delegation. The registrar's publication process, cached NS records and any nameserver address (glue) records all affect the transition. Long delegation TTLs can make a nameserver change take longer to become visible than an ordinary record change. Check the provider's stated publication time and the actual record TTLs when planning it.

Applications and proxies — Browsers, applications and proxies may maintain their own caches. X4B HTTP hostname backends have a separate cache lifetime to include in your migration plan.

Plan and check a change

  1. Check existing records and TTLs before scheduling the migration.
  2. Lower the relevant TTL in advance where supported, then wait for the previous TTL to elapse.
  3. Configure and test the new destination before changing DNS.
  4. Check that the authoritative nameservers publish the new answer, then compare several recursive resolvers and an affected client's answer.
  5. Keep the old destination available while cached answers and existing connections drain.

For nameserver changes, check the parent delegation and required glue as well. An incorrect record or unreachable nameserver needs correction; waiting alone will not resolve it.

See negative caching and stale-answer handling for the underlying DNS behavior.