If v2rayN fails to start or websites won't load, first check whether the core loaded its config and started listening locally, then inspect the outbound connection. Using v2rayN with the Xray core on Windows as an example, this guide shows how to capture logs for a failure and narrow down the cause using addresses, ports, and error context.
Where to find v2rayN logs and choose a log level
Start with the “Information” area in the main window and look for output around the time you switch nodes or start the core. To set the log level, go to “Settings” → “Parameter settings,” then find “Log level” under the Core settings. Group names may vary between versions. With a custom config, also check whether the loaded config specifies its own log output location and level.
Record your current setup
Note the v2rayN version, the core used by the node, and the local proxy port. Avoid updating subscriptions, switching nodes, and changing ports all at once; even if the connection recovers, you won't know which change fixed it.
Set the level to info
Temporarily set the core log level to
info, save, and restart the current core. The corresponding field in an Xray config islog.loglevel.warninganderrorproduce less output and are suitable for everyday use, but may not show each step a request takes.Reproduce the problem once
Note when you make the request, then retry once in the app that originally showed the error. If the issue occurs at startup, restart only the core; there's no need to keep refreshing the page. Keep the failing line and a few lines around it rather than copying only the last line in the window.
Capture more detail, then restore the setting
If
infodoesn't show where the failure occurs, switch todebugtemporarily and retry once, then restore the original level. Before sharing logs, redact subscription URLs, credentials, private domain names, and public IP addresses. Keep the error keywords, ports, and identifiers that link entries from the same request.
Reading log structure: local inbound, routing, and remote outbound
A log may include client messages, core output, and speed-test results. To read core output, first match its timestamp to the operation, then check the component and the action that failed. Some core logs chain multiple errors with >: the left side often summarizes the failed operation, while the right side gives a more specific system error. Compare entries in the same error chain; don't combine nearby requests into a single cause.
| Log entry | What it tells you | What to check next |
|---|---|---|
listen tcp 127.0.0.1:10808 | This points to a local listen operation | Check the same line for a bind failure; don't test the server port |
accepted | The core received this inbound request | If an inbound-to-outbound tag appears next, check that it matches the expected route |
dial tcp | A component is attempting to open a TCP connection | Use the destination address and port in the log to tell whether it's a node, a direct destination, or a local service |
Diagnosis: check the route for the same request
If the browser fails and the corresponding accepted entry is present, but the outbound tag shows a direct route, check that route rule and its order first. Reinstalling the client or repeatedly switching nodes won't change the outbound route selected for this request.
Core startup failures: port conflicts and config parsing errors
When an error occurs during startup, a browser proxy error is usually just a downstream symptom. First confirm that the core loaded its config successfully, then check whether it can listen on the local port. Both failures can prevent requests from reaching the core, but they require different fixes.
Error: Only one usage of each socket address (protocol/network address/port) is normally permitted.
Cause and fix: Windows reports this bind error when the current process can't use that address and port because they're already in use. Check which process is listening and whether it's another client instance or a leftover core process. Close the conflicting program cleanly, or change the local port and update the proxy settings in your apps.
Error: invalid character '}' looking for beginning of object key string
Cause and fix: The parser found a closing brace where it expected the next object key, often because of a trailing comma after the last field. Use the filename, line and column, or offset in the error to find the config that was actually loaded. Fix the syntax, then check the field types and core support.
If the listening port in the error is 10808, you can check it in Windows PowerShell. If the log shows a different port, replace the number in the command. If no listener is found, check the error timestamp again; the process causing the conflict may have exited.
Get-NetTCPConnection -LocalPort 10808 -State Listen |
Select-Object LocalAddress, LocalPort, OwningProcess
OwningProcess is the process ID; look it up on Task Manager's “Details” tab. If you change the local proxy port, reapply v2rayN's system proxy settings and check any apps where you entered a proxy address manually. If an auto-generated config fails to parse, correct the source in the node editor first; edits to a temporary config may be overwritten the next time it starts.
Connection timeouts, refused connections, and TLS handshake failures
Once the core has received a request, determine whether dialing failed or whether the TCP connection succeeded and the failure occurred during the handshake. These have different things to check: for the former, look at the actual destination address and port; for the latter, check settings such as the security type, server name, and certificate.
Error: i/o timeout
Cause and fix: A network operation exceeded its time limit, but this message alone doesn't tell you which layer timed out. If it follows dial tcp, test the destination TCP port shown in the log. If the error occurred while reading a response or during the handshake, check the surrounding entries in the same error chain.
Error: connection refused
Cause and fix: The connection was actively refused, commonly because no service is listening on the destination port or an intermediary rejected it. Check whether the address is local. For a local address, check the listening process; for a remote one, verify the node port and ask the service operator to confirm that it's listening and allowed through.
Error: x509: certificate signed by unknown authority
Cause and fix: Standard TLS certificate verification can't chain the certificate it received to a trusted root. First confirm that you're connecting to the expected service, then ask the service operator to check the certificate chain. On a managed network, ask the administrator about its trust requirements; don't bypass verification just to silence the error.
Error: remote error: tls: handshake failure
Cause and fix: The remote server returned a handshake failure alert. This line alone doesn't reveal which setting is mismatched. Check the node's security type, server name, and port. If the config uses REALITY, verify the server name, public key, and short ID against the parameters provided by the server. Don't apply standard TLS certificate settings to REALITY.
For a node that uses TCP, you can test the server endpoint separately in PowerShell. When prompted, enter the node's server address and remote port—not the subscription URL, TLS server name, or local proxy port.
$serverAddress = Read-Host 'Node server address'
$serverPort = [int](Read-Host 'Node remote port')
Test-NetConnection -ComputerName $serverAddress -Port $serverPort
Common questions about troubleshooting logs
Websites won't load, but there's no new output in the Information area?
Temporarily set the log level to info and try again. If you still can't find a matching inbound entry, confirm that the core is listening, then check v2rayN's system proxy status and Windows “Settings” → “Network & internet” → “Proxy.” If the app has its own proxy settings, make sure its address, port, and proxy protocol match the actual inbound configuration.
The log only says failed to find an available destination. Does that mean it's DNS?
No. This is a general message indicating that an outbound connection couldn't be established; it doesn't necessarily mean domain resolution failed. Look further along for the specific error: check the domain spelling and resolution if you see lookup and no such host. If the destination IP and a dial timeout are already present, first check the connection to that IP and port.
The node works. Why is the subscription update still timing out?
Subscription updates connect to the subscription server; a working node doesn't mean the subscription server is reachable. In the main window, open “Subscription groups” and compare the update options for “Direct” and “Through proxy” to identify which one fails. Before trying an update through the proxy, verify that the current node can load a webpage. Before sharing update logs, remove the path and query parameters from the subscription URL.
The log occasionally shows EOF. Should I switch nodes?
First check whether the web request actually failed. EOF means the end of the stream was reached; on its own, it doesn't tell you who closed the connection or why. If the same operation repeatedly triggers unexpected EOF during the handshake, check the remote port and transport settings, then verify with the same test request.