Use this guide when Shadowrocket shows node latency results but your browser says it can’t find the server or keeps loading. Note your current network, Global Routing mode, and active configuration first. Then distinguish DNS lookup failures from unreachable resolved addresses and later TLS connection failures. Change one setting at a time and retest the same domain. This guide assumes you already have your own service details; it does not provide, sell, or recommend nodes or subscriptions.
1. Check your device and distinguish latency tests from web access
If you need to install or reinstall the app, start at the download page for all platforms and select your device: iPhone and iPad versions are available on the App Store for a one-time payment of $2.99. The developer is Shadow Launch Technology Limited; check the App Store listing for system requirements. Versions for Mac (Apple Silicon and Intel), Windows, Android, and Linux are free. Don’t make changing the installation package your first DNS troubleshooting step. Keep your current configuration and error details.
A latency test only confirms that its test target responded using that particular test method. A TCP connection test and a connectivity test to a specific URL cover different parts of the connection. Even if a test includes a DNS lookup, it doesn’t prove that every website domain, record type, or subsequent connection is working.
This sequence helps divide troubleshooting into stages; it doesn’t mean every request follows exactly the same internal order. Domain rules can be evaluated before an IP address is available, and the proxy server may resolve the destination domain. Page images and scripts can also request other domains, so a page that displays text but no images may indicate a failure with just one resource domain.
- Server not found: Check the domain spelling, DNS response status, and whether the resolver is reachable.
- Long wait followed by a timeout: The DNS query may have gone unanswered, or the IP address may have been resolved but the connection to the destination failed.
- Certificate or secure connection error: Check the system clock, TLS, and destination service. This symptom alone doesn’t establish a DNS problem.
2. Follow a consistent process and keep reproducible test notes
The instructions below use Home, Config, and Settings as interface labels. The exact control locations vary by platform; refer to the current interface. Record the original values first. Don’t change the node, DNS, and rules all at once, or you won’t know which change restored access.
-
Record the current state
On Home, note the selected service and Global Routing mode. In Config, note the active configuration name. Record the failing domain, time of the failure, whether you’re on Wi-Fi or cellular, and the browser’s exact error message.
-
Choose fixed test domains
From your own browsing history, choose one domain that consistently fails and one known to work as a control. Close existing tabs and open the domains again so a cached page isn’t mistaken for a fix.
-
Check the active configuration
Open the active configuration in Config and check the DNS settings in [General]. Then go to Settings → DNS and review the global DNS options. If both contain settings, first confirm which values are actually in use. Don’t infer the final behavior from a single field on screen.
-
Control automatic connections
Check Settings → On Demand. If automatic connections are enabled, note the current setting before disconnecting for comparison, so the tunnel doesn’t reconnect during testing. Restore the original setting when you’re done.
-
Change one setting and retest
Change just one DNS address or related option, reconnect, and try the same domain again. Record the result before deciding whether to keep or revert the change. If needed, switch to another network for comparison.
3. Identify the failure stage from the DNS response
If diagnostic output or logs include a DNS response code, save the exact text. The statuses below are common in the DNS protocol and lookup tools; they may not appear identically in Shadowrocket on every platform. A browser message such as “This site can’t be reached” isn’t enough to determine the response code.
DNS response: NXDOMAIN
Cause and fix: The resolver says the queried name doesn’t exist. Check the full domain and spelling, then verify it through another reachable DNS path. Changing the proxy port won’t fix a name that genuinely doesn’t exist.
DNS response: SERVFAIL
Cause and fix: The resolver couldn’t complete the query. The issue may involve an upstream resolver, authoritative server, or DNSSEC validation. First check whether other domains resolve, then test the same name with another resolver and save the response status for further diagnosis.
DNS response: REFUSED
Cause and fix: The server refused the query, often because of recursive query permissions or access policies. Confirm that the DNS service allows requests from your current network instead of repeatedly increasing the wait time.
Query status: timeout
Cause and fix: No usable response arrived before the timeout. This doesn’t mean the domain doesn’t exist. Check the resolver address, transport protocol, port, and outbound route, then test again on another network.
Another case that’s easy to misread is NOERROR with no answer for the queried record type. For example, a domain might have an A record but no AAAA record; an empty AAAA answer doesn’t mean the domain doesn’t exist. Record the query name, record type, and response content—not just the status code.
4. Understand DNS addresses, transport protocols, and ports
DNS settings involve more than a server address: they also determine the transport method and whether the server is reachable over your current network. In [General], dns-server specifies the resolver. Use an address format supported by the current platform and setting. Don’t enter a DoH URL as a plain IP address or swap ports between protocols arbitrarily.
Takeaway: Check the transport path before changing resolvers
If traditional DNS queries time out on a network but correctly configured DoH queries succeed, check the traditional DNS transport path first. This doesn’t by itself prove there’s a problem with the domain’s records or the proxy service.
| Setting | What it means | What to check during troubleshooting |
|---|---|---|
| System DNS | Uses the DNS resolution settings provided by the current network. | Useful as a control. Corporate, campus, and intranet domains may rely on that network’s resolver. |
| Custom resolver | Specifies the server used for queries. | Use a service address you’re authorized to use and can reach from your current network. Don’t copy configuration from an unknown source. |
| DoH | DNS messages are sent over HTTPS; the endpoint includes a protocol, host, and path. | Check the complete URL provided by the service, certificate requirements, and network reachability—not just the hostname. |
| DoT | DNS messages are sent over TLS. | Confirm that the relevant port is reachable and check the server name required for certificate validation. |
| A / AAAA | Returns IPv4 and IPv6 addresses separately. | Check actual support on the local and outbound paths. An AAAA record doesn’t mean IPv6 connectivity is available. |
If an encrypted DNS endpoint uses a domain name, its address must be resolved before a connection can be established. This initial lookup is called bootstrap resolution. If the path it depends on is already broken, switching to a new resolver URL that uses a domain name may not help. Check the initial resolution path; don’t disable certificate validation to work around the issue.
5. Check Global Routing, rules, and IPv6
In Home → Global Routing, Proxy, Direct, Config, and Scene describe different overall traffic handling modes: Proxy generally routes traffic through a proxy, Direct connects directly, Config follows the configuration, and Scene is tied to scene settings. Record the current mode while troubleshooting. These modes don’t mean that every DNS request automatically takes exactly the same route.
- Fails in Config but works in Proxy: Check which rules match and which outbound route they select. This suggests the two tests took different paths; it doesn’t by itself prove the DNS server is at fault.
- Fails in Direct: The local network may not reach the destination, or its system DNS settings may differ. A failed direct connection doesn’t prove that your existing service is unavailable.
- Fails on only one network: Compare the DNS and IPv6 connectivity provided by Wi-Fi and cellular, and check whether you need to complete network sign-in first.
- Only intranet domains fail: Check whether they require an internal DNS resolver. Sending every query to an external resolver may prevent internal names from resolving.
DOMAIN-SUFFIX matches a domain suffix, IP-CIDR matches an IP range, GEOIP matches against an IP geolocation database, and FINAL handles traffic that didn’t match an earlier rule. Domain and IP rules use different information. Some IP-based rules may require a resolved address, so DNS and rules can be related.
When reviewing an existing configuration, use DOMAIN-SUFFIX,example.com,DIRECT to understand the syntax for sending a domain suffix directly. This is an example for reading rules, not a request to add it to your active configuration. Domain exceptions usually need to come before broader rules that would match first. Also check for outdated mappings in [Host] that could keep sending requests to an old address.
6. Check subscription updates, caches, and troubleshooting notes
A failed subscription update and a failed lookup for a website can involve different domains. If your service node uses a domain as its server address, distinguish between resolving the proxy server’s address and resolving the destination website’s address. If the former fails, the tunnel may never connect; if the latter fails, an existing connection or some tests may still work.
After updating your own subscription, return to Home and confirm the currently selected service and its settings. Don’t stop at seeing the refresh complete. A subscription may update server addresses and ports, but that doesn’t mean DNS, rules, or [Host] in the active Config were also changed. Checking the effective configuration is more useful than importing it repeatedly.
- Record the minimum steps to reproduce: The failing domain, whether the query was for A or AAAA, the response status, time of the attempt, and whether you were on Wi-Fi or cellular.
- Record the traffic settings: Current Global Routing mode, active configuration, DNS transport method, and whether the browser’s own Secure DNS is enabled.
- Record comparison results: How the same domain behaved on another network, through another DNS path, or after restoring the original settings. Clearly mark any test you haven’t performed as not tested.
- Remove access credentials: When contacting your own service provider, redact full subscription URLs, passwords, tokens, and other authentication details. Include only the error context needed to troubleshoot.
If Secure DNS is enabled in the browser, DNS lookups may not rely entirely on the system or Shadowrocket DNS settings. If changes in the app don’t affect the result, include the browser’s DNS settings in your comparison instead of adding more resolvers. Managed company devices may also be subject to policy restrictions; follow your device-use requirements.
Takeaway: Finish by checking the complete results for the same domain
You can pinpoint whether the problem is with DNS, rules, or the destination connection only when your notes include the resolution result, actual outbound route, and webpage connection behavior. If DNS returns a usable address but TLS still fails, stop switching DNS servers and check the system clock, certificates, and subsequent connection path.
Download Shadowrocket and view setup guides
The iPhone / iPad version costs a one-time $2.99 on the App Store. Versions for Mac / Windows / Android / Linux are free. Mac downloads are available for Apple Silicon and Intel; check the App Store listing for iPhone / iPad system requirements. Get the app for your device, then follow the guide to add your own service details.