How to Test IPv6 Addresses? A Comprehensive Troubleshooting Guide—From Online Self-Checks to Command-Line Methods
By 2026, the adoption of IPv6 is no longer merely a "forecasted trend." Major cloud providers now assign IPv6 addresses by default, mobile networks inherently prioritize IPv6 traffic, and some applications have even begun to listen exclusively on IPv6. For operations and maintenance (O&M) personnel, however, a more practical question arises: once an IPv6 address is assigned, how can one verify that it is truly functional, the link is active, and applications are responding as expected?
Successfully "pinging" an address does not guarantee that TCP-based services will function; likewise, the mere presence of an IPv6 address does not confirm that routing and firewall configurations are correct. IPv6 testing must be conducted in layers—ranging from quick self-checks via a browser to systematic, layer-by-layer troubleshooting from the command line—with each stage addressing specific concerns.
Step 1: Browser-based self-check—verifying basic availability
Before diving into command-line diagnostics, perform a quick "health check" using an online tool to determine whether your current network environment possesses basic IPv6 capabilities.
`test-ipv6.com` is the most widely recognized site for this purpose. Upon visiting, it automatically runs a suite of tests to evaluate whether your browser, operating system, and network connection fully support IPv6. It then provides a score (from 0 to 10) and details the status of each individual test. If the score is low, the page will explicitly identify the point of failure—for example, "no IPv6 address assigned," "DNS resolution does not support AAAA records," or "unable to connect to IPv6 targets."
Another useful resource is `ipv6.icanhazip.com`. By entering `https://ipv6.icanhazip.com` into your browser's address bar, you can verify your connectivity: if the site returns a colon-separated hexadecimal address, your network egress point has functional IPv6 connectivity. If it returns an IPv4 address or the connection fails, it indicates that the IPv6 link is not yet properly established.
The value of this step lies in its ability to quickly distinguish between two types of issues: a complete lack of IPv6 capability versus the presence of an IPv6 address that is incorrectly configured. The former requires troubleshooting at the network access layer, while the latter is the focus of subsequent command-line diagnostic procedures. Step 2: Layered Command-Line Testing—From Addressing to Application
If an online self-test indicates that an "IPv6 address is present but the connection is abnormal," or if you need to verify your IPv6 configuration directly on a server, command-line tools are the only way to pinpoint exactly which layer is causing the issue.
Layer 1: Verify the assigned address. Run `ip -6 addr show` (or check the IPv6 section of `ifconfig`) to see if there is a global IPv6 address—one that does *not* begin with `fe80::`. Addresses starting with `fe80` are "link-local" addresses; they are only valid within the local Layer 2 network and cannot be used for internet communication. You must have an address with a `scope global` designation (e.g., `2400:xxxx::xxxx`) to communicate over the public internet. Additionally, use `ip -6 route show` to confirm the existence of a `default via` route pointing to the correct IPv6 gateway.
Layer 2: Test Layer 3 (Network Layer) connectivity. Use `ping6` (or `ping -6`) to probe a known, accessible public IPv6 address—such as Cloudflare’s `2606:4700:4700::1111` or Google’s `2001:4860:4860::8888`. If you receive a response, the IPv6 route from your machine to the target network is functional. If the request times out, do not immediately conclude that "IPv6 is broken"—some networks restrict ICMPv6 echo requests. In such cases, you should evaluate the situation by performing the next layer of testing.
Layer 3: Test the routing path. If `ping6` times out or shows abnormal latency, use `traceroute6` (or `tracert -6`) to trace the path. Observe at which hop the data packets are dropped, or check for inefficient, circuitous routing. A common issue with IPv6 routing is incorrect source address selection; when a server has both IPv4 and IPv6, applications might take an inefficient path if the egress interface is not explicitly specified.
Layer 4: Test the application layer (the most critical layer). Connectivity at Layer 3 does not guarantee that TCP or HTTP services are functional. Use `curl -6` to force a connection to an external service via IPv6:
curl -6 -sS -o /dev/null -w "%{http_code}\n" https://[2606:4700:4700::1111]/
The `-6` flag forces `curl` to use only the IPv6 protocol stack. If an HTTP status code like `200` or `301` is returned, it indicates that the entire path—from addressing and routing to the TCP handshake and TLS negotiation—is functional. If `ping6` succeeds but `curl -6` times out, the issue likely lies with a firewall blocking IPv6 TCP traffic, or the application itself may be configured to listen only on an IPv4 address.
Step 3: Distinguishing between "the host has IPv6" and "the website supports IPv6"
A server’s ability to initiate outbound connections via IPv6 is distinct from an external user’s ability to access your website via IPv6.
To verify whether a website truly supports IPv6 access, check the domain's AAAA records and confirm that the application is listening on an IPv6 address. Run `dig AAAA yourdomain.com +short` to ensure the correct IPv6 address is returned. Then, check the listening status of your web service: `ss -6 -tlnp | grep :443` should show `[::]:443` or a specific IPv6 address in the `LISTEN` state. If you only see `0.0.0.0:443`, the application is listening exclusively on IPv4; in this case, even if an IPv6 user resolves the domain to an IPv6 address, they will be unable to establish a connection.
A more intuitive way to verify this is to check the "Remote Address" field for a request within the Network panel of your browser's developer tools (F12). If it displays a colon-separated hexadecimal address (e.g., `240e:390:2c0:1::2`), the request is indeed using IPv6. If it shows a dot-decimal IPv4 address, the connection is still using IPv4.
Common Failure Scenarios and Troubleshooting
Scenario 1: `ping6` succeeds, but `curl -6` fails. The most common cause is that the cloud platform’s security groups or server firewalls are configured only with IPv4 rules. Since IPv4 and IPv6 firewalls operate as separate sets of rules, you must ensure that the target ports are explicitly opened in both the security groups and the server-level firewalls (such as `ip6tables` or `ufw`).
Scenario 2: The host has a global IPv6 address but cannot access the external network. Check whether the default route points to the correct gateway. If a default route exists but is associated with the wrong network interface, or if the gateway address does not match the prefix assigned to the host, the data packets will be dropped.
Scenario 3: DNS resolution returns an AAAA record, yet the service remains inaccessible. Verify whether the web service is actually listening on an IPv6 address. For instance, in Nginx, a configuration of `listen 443 ssl;` only listens for IPv4; you must explicitly add `listen [::]:443 ssl;` to accept IPv6 connections. This is the most frequent configuration oversight in dual-stack deployments.
From "Testable" to "Consistently Available"
A one-off test only answers the question: "Is it working right now?" In a production environment, the stability of an IPv6 link requires continuous monitoring. Long-duration `ping6` tests (e.g., `ping6 -c 1000`) can help identify periodic fluctuations in packet loss, while the IPv6 mode of `mtr` (`mtr -6`) allows for ongoing tracking of path quality. By integrating IPv6 connectivity checks into your existing monitoring framework—rather than relying on manual troubleshooting after receiving user complaints—you can effectively transform the mere "existence of an address" into "service availability."
For services requiring a dual-stack deployment, it is advisable to confirm the IPv6 support status of your chosen product—either via a support ticket or the product documentation—before purchase. Once you have been assigned an IPv6 address, follow the layered verification approach outlined above, covering everything from address assignment and routing to application-level listening. With independent IPv4 addresses provided as a standard feature across the entire product line, backed by a stable CN2 GIA network infrastructure, you gain reliable connectivity for the IPv4 component of your dual-stack environment. This allows for the testing and rollout of IPv6 to proceed incrementally, without disrupting existing operations. Furthermore, a consistent pricing policy for renewals ensures that the long-term costs of maintaining a dual-stack architecture remain predictable.
CN
EN