If a connection fails or a page won’t load, reproduce the issue once, then use the timestamp, addresses, and surrounding log entries to find where it failed. This guide is for readers who can open their client but aren’t sure whether to check the local proxy, remote connection, or TLS settings. By the end, you’ll know which setting to check first in a log excerpt.
Set the Log Level and Reproduce the Issue
Temporarily use info to capture a complete attempt; switch to debug if you need more detail. warning is usually enough for everyday use. error records only errors, so it can miss useful clues from earlier in the process. The core log level and access logging are separate settings: increasing the detail level does not necessarily show every access record.
Find the Core Logs
In a typical v2rayN 6.x interface, open “Settings” → “Parameter Settings” → “Core: Basic Settings” and find “Log Level.” Save your changes, restart the core or disconnect and reconnect, then check the log panel on the main screen. In v2rayNG, open the menu in the top-left corner and select “Log.” Menu names may vary by version; make sure you’re looking at connection logs, not speed test results.
Reproduce It Once
Note your local time, keep the current node and network unchanged, and stop any batch speed tests. Then use the same app that had the issue to visit the same destination once. If your browser has many tabs open, close the unrelated ones first so background requests don’t bury the failure among a string of timeouts.
Keep the Lines Before the Error
Starting from the time the request was made, keep the entries for that same request: the inbound connection, the selected outbound, and the error. Before sharing, redact subscription URLs, account identifiers, destinations, and internal addresses. Use consistent labels for your device and the remote host, and preserve the original error text and the order of the entries.
Change One Setting at a Time
Change just one setting based on where the error occurred, then test again using the same app and destination. For example, correct the proxy port in the app first; don’t also switch nodes, DNS, and transport settings. Once you’ve confirmed the result, restore the log level you were using before troubleshooting.
Log Structure: Timestamps, Addresses, and Outbound Tags
First, distinguish the addresses in the log: the app connects to the local proxy listener, the request specifies the destination website, and the proxy’s outbound connection goes to the node server. When a proxy is involved, these three addresses may differ. The same connection refused error can point to different checks depending on where it appears.
The two lines below illustrate log fields; the addresses are examples. The first records a local request being accepted, and the second records a timeout while the core connects to a remote host. Even if both lines appear in the same second, use the connection ID (if available), addresses, and surrounding entries to confirm whether they belong to the same request.
2026/10/09 10:14:28 from tcp:127.0.0.1:53162 accepted tcp:example.org:443 [socks -> proxy]
2026/10/09 10:14:28 [Warning] dial tcp 203.0.113.10:443: i/o timeout
53162is the example’s temporary source port on the app side, not the proxy listening port that should be entered in the app.acceptedmeans the core received the request; it doesn’t mean the website returned a response. If access logging is enabled but there’s no inbound entry, check whether the app uses the system proxy and whether its local port is correct.socksandproxyare inbound and outbound tags in the example configuration, not fixed labels used by every client. If the request actually selected a direct outbound, check the routing rules first instead of changing the node’s TLS settings.
Common Errors and What to Check First
The final log entry after retries may only summarize the failure. If you see failed to find an available destination, look earlier in the same request for lookup, dial, read, or handshake. This summary alone isn’t enough to identify a DNS problem.
Error: connection refused
What it means: The connection attempt was rejected. Often, nothing is listening on the specified port, though a network device may also actively reject the connection. Check first: If the address is 127.0.0.1, compare the proxy port in the app with the client’s listening port, and check that the core started successfully. If it’s a node address, compare the port with the subscription; for a self-managed server, check that a service is listening on that port.
Error: i/o timeout
What it means: A network operation took longer than the allowed wait time. This message alone doesn’t mean the node is offline. Check first: If the log starts with dial tcp, check the address and port it’s trying to connect to, then test the same node on another working network. If it says read tcp, the timeout happened while reading; check the preceding entries to see whether it was waiting for a handshake or a response.
Error: remote error: tls: handshake failure
What it means: The remote peer sent a failure alert during the TLS handshake, but this message doesn’t identify which setting is mismatched. Check first: If this is a node connection, verify that the node port, security type, and serverName (SNI) match the service instructions. For REALITY, compare against the original configuration; don’t treat it as a regular TLS node and fill in different values.
Error: x509: certificate has expired or is not yet valid
What it means: The certificate isn’t valid for the device’s current time. Check first: Enable automatic system time and correct the date, then test again. If the time is correct and the error persists, ask the service maintainer to check the certificate actually served on that port. Don’t disable certificate verification to hide the problem.
Error: no such host
What it means: The name shown in the log’s lookup entry couldn’t be resolved to an address. Check first: Determine whether the name is the node’s domain or the destination domain, then check its spelling and the DNS resolver responsible for it. Don’t casually replace the domain with an IP address: even if the connection succeeds, this may affect TLS name verification or service matching.
Choose Your Next Step Based on the Retest
If the same node works on another network, compare the DNS results or connection behavior on each network first. If only one app fails on the same network, check the inbound and outbound paths used by that app. A single comparison can narrow down the cause, but it can’t prove that the server is at fault.
Speed Test Passes, but Why Won’t Websites Load?
A speed test may target a different destination from the one your browser is trying to reach. Reopen the page that failed and check the log entries at the corresponding time for its destination and outbound tag. If there’s no matching entry, check the browser’s proxy settings first. If there is an entry but it uses a direct outbound, check the active routing rules.
The Client Reports a Port Conflict on Connect. Should I Change the Node Port?
First, check for a startup error such as bind: address already in use. If the log shows a local listening address, check for another instance of the client or another program using that port. If you do need to change the local listening port, update any apps where you entered it manually. Don’t change the node port.
Only One EOF Entry: What Should I Check?
EOF means the stream ended; by itself, it doesn’t explain why. Reproduce the issue once and check earlier entries for the same connection. If it consistently stops during the handshake, compare the port and transport settings with the server’s logs from the same time. If it only appears occasionally when you close a page and everything works, leave the settings alone for now.
Disconnects Only When Android Goes into the Background: What Should I Check?
First, compare the log timestamps before and after the screen turns off. If logging stops and then shows the core starting or the VPN being re-established, check the background activity restrictions under the system’s “Settings” → “Apps” → “v2rayNG” battery settings. The path varies by device. If the client keeps running, check whether Wi-Fi switched to mobile data before the disconnect, then look for a new dial timeout.