Category: Network connectivity troubleshooting

So you have just found out that you no longer have access to your VPS. Let’s find out what is causing the issue. We will be using ping.pe for our troubleshooting.

Normal output

Here is what a normal output of ping.pe looks like:

The above output shows that the VPS is accessible from all parts of the world. If you are seeing this output however you are still unable to connect to your VPS, then most likely there is a software configuration issue which you can solve by using the Interactive Console in KiwiVM.

No connectivity at all, complete blackout

The output below shows 100% packet loss from all locations, meaning your VPS is completely inaccessible:

First of all, use KiwiVM to verify that your VPS is actually running.

Next step, verify your firewall configuration. The following article explains how to do that:

If there are no issues with firewall, then most likely there is an issue with the network configuration inside of your VPS. The easiest way to solve these issues would be to perform an OS reinstall. You can also use the Interactive Console in KiwiVM to troubleshoot such issues, however specific steps fall beyond the scope of this article.

Slow connectivity, packet loss

Here’s what packet loss looks like:

In this particular example we can see that there is heavy packet loss to 8.8.8.8 from all China locations (except one). To further diagnose the issue open up the full mtr report:

From this report it is clear that the issue is within China. If in doubt, we recommend performing mtr on the return path (in opposite direction – from the affected IP back to your home IP). Due to path asymmetry you may find a bit different output which may clarify the issue.

In many cases connectivity issues may arise from incorrect iptables configuration. Very often misconfigured iptables is the result of a recent change of the IP address of the VPS due to migration performed via KiwiVM control panel. In cases like these you become completely locked out of your VPS, and ping.pe output looks like this:

Notice that VPS is completely unreachable. If you open up the MTR report, you will see the following:

As you can see from the above example, the physical node “v403.sioru.com” is functioning fine, however the specific VPS we are pinging is offline.

The quickest way to verify that iptables is indeed the issue is to use Interactive console in KiwiVM to stop iptables service (please search online on how to stop iptables, the exact command depends on your distribution). If network connectivity is restored, then we know that iptables is the issue.

To check your iptables rules, issue the following command:

iptables -L -n; iptables -L -n -t nat

Note that in order for this command to work properly, your iptables service must be restarted again.

If you see old VPS IP in the output, especially in the NAT table, then you will have to manually modify the rule (again, search online on how to delete and add iptables rules).

Alternatively, you can wipe all iptables rules with the following commands:

iptables -F; iptables -t nat -F; iptables-save > /etc/sysconfig/iptables

As the last resort, you could always reload your VPS by installing a fresh operating system, this will completely restore all network connectivity.