Preview Your Website on a New Server Before Changing Public DNS

You can test your own website on a replacement server before directing public visitors there. The key is to send an authorized test request to the new address while retaining the intended hostname. That lets you inspect the site’s configuration without using a public DNS change as the first experiment.
A Windows RDP can provide a dedicated testing workspace for this process. Use it only for domains and servers you own or are authorized to test. Obtain the destination address from the hosting administrator; do not search for hidden origins belonging to another organization or bypass access restrictions.
Prepare the replacement site before making requests
Confirm the intended hostname, approved replacement IP, HTTPS configuration, and a way to identify the new deployment. A harmless build marker or a corresponding access-log entry can help distinguish the replacement from the current production server. Keep public DNS unchanged during the initial checks.
Decide how the test copy handles email, payments, webhooks, scheduled jobs, and customer data. A cloned website can contain production credentials or point to live services. Use approved test integrations and dummy records where appropriate. Do not assume a page is isolated merely because ordinary visitors cannot yet reach its new address.
Use a request-scoped address override
With an approved curl installation available inside Windows, a basic template is curl.exe --resolve "www.example.com:443:203.0.113.10" --head "https://www.example.com/". The hostname and address here are documentation examples, not a live test target. Replace them with your own authorized values.
The official curl reference describes --resolve as an address override for a specific hostname and port. Unlike changing public DNS, the option applies to the curl operation. Keep the intended hostname in the URL rather than substituting the bare server IP.
This example assumes that curl is permitted to connect through the intended test path. A required proxy or gateway can affect the connection and name-resolution behavior. Do not bypass an organizational control to make the command work; ask the administrator to provide an approved route and verify the destination in the new server’s logs.
Keep certificate verification enabled
The replacement server should present a certificate appropriate to the hostname being tested. Do not add an insecure option simply because validation fails. Check whether the certificate, hostname coverage, trust configuration, and system time are correct. A test that ignores the failure does not prove the production HTTPS configuration is ready.
Curl’s TLS certificate verification documentation explains the default verification behavior and supported trust configuration. For an intentionally private test environment, use the organization’s approved certificate authority and trust process. Do not install an unknown root certificate from a support chat.
Read redirects before following them
The example requests response headers so you can inspect the status and any Location header. If the response redirects from the www hostname to the apex domain, your first override does not automatically cover that second hostname. Confirm the complete intended route before adding a second explicit mapping for a domain you control.
A HEAD request is not a full browser test, and some applications handle it differently from GET. Use an authorized ordinary page request when needed and compare the resulting content with the new deployment’s marker or logs. Do not interpret one 200 response as proof that every asset, application route, and integration is correct.
Use an acceptance checklist for the actual website
| Area | What to verify |
|---|---|
| Server identity | The request reached the intended replacement deployment |
| HTTPS and redirects | The expected hostname remains valid throughout the route |
| Pages and assets | Representative pages, images, styles, and scripts load from intended locations |
| Application data | The test uses the correct database and approved sample records |
| Outbound integrations | Tests cannot create unintended real payments, messages, or webhook actions |
| Cutover readiness | The final data-sync and rollback responsibilities are documented |
For interactive browser testing, use the hosting platform’s approved preview method or an administrator-reviewed local mapping in a dedicated test environment. Confirm what hostname the browser actually uses and which deployment receives its requests. A preview URL with a different hostname may not reproduce every cookie, redirect, or certificate behavior of the production hostname.
Remove temporary overrides and separate preview from cutover
A curl command-line override ends with that operation, but a persistent local hosts-file mapping must be removed deliberately when testing finishes. Keep a record of any persistent change and restore only the entries you changed. Verify a normal, non-overridden request afterward so that later work does not accidentally continue against the test server.
Plan the public DNS change separately, including the final data synchronization, write ownership, rollback criteria, and monitoring. A preview validates selected behavior before cutover; it does not transfer recent production orders or guarantee uninterrupted service on its own. Review the DNS troubleshooting guide when the observed lookup differs from the expected public record.
For ClovRDP-side connectivity questions, contact support with the approved destination and redacted error. The website’s hosting and application administrators should own the migration decision and the production cutover.
Featured photograph: Justin Morgan on Unsplash. Technical references reviewed on 17 October 2026.


