Migration · By Rivtor · Updated
Move a custom domain between hosts with a controlled cutover
Prepare the destination hostname, TLS and application configuration before changing DNS. Inventory every affected record, preserve email settings and define the data-write and rollback boundaries. DNS changes route requests; they do not move application data or active sessions automatically.
Compare Rivtor plans for your workload →
Inventory records and application dependencies
Record the current A, AAAA, CNAME, TXT and mail-related records relevant to the domain. A website move should not accidentally remove MX, SPF, DKIM or verification entries. Identify whether a CDN or reverse proxy sits in front of the origin and where TLS is terminated.
Also list application references: allowed hostnames, CORS origins, OAuth callback URLs, webhook endpoints and canonical URLs. A DNS record can point correctly while the application still rejects the hostname or redirects users back to the old location.
Prepare the destination before routing visitors
| Readiness item | What to verify |
|---|---|
| Domain registration on host | Ownership validation and correct project |
| TLS | Certificate can be issued for the intended hostname |
| Application | Health and representative routes work |
| Data | Final transfer and write boundary are defined |
| SEO | Canonical hostname and redirects are consistent |
Use the destination’s documented domain instructions. On Rivtor, start with the CLI and domain reference and verify the current required records. Do not copy a provider’s IP address from an old blog post.
Schedule the record change
If lowering a TTL, do it sufficiently ahead for the old TTL to expire. Caches and resolvers do not all update at the instant you save a record. Keep the previous origin available while observing the transition, and avoid serving conflicting writable application states.
Do not change nameservers unless that is part of the intended move. A simple host cutover often requires a record update rather than replacing the whole DNS service. Follow the actual provider setup for proxied records and certificate validation.
Verify traffic and rollback assumptions
Check DNS resolution from more than one resolver, HTTPS status, redirects, cookies, browser requests and the application’s critical synthetic transaction. Confirm the public canonical URL uses the intended hostname. Watch origin responses during the transition.
If the new environment has accepted writes, routing traffic back is a data-recovery decision too. Preserve new records deliberately rather than treating DNS rollback as a complete application rollback. Keep a dated record of old and new configuration for the release owner.