Viscosity DNS server stops working

Hi,
We have rolled out Viscosity to our entire engineering department using a 30-seat license and seeing some have issues where the Viscosity DNS server stops responding and this causes all DNS queries to timeout and slows down internet browsing etc dramatically!
The only solution so far has been to disconnect all connections and reconnect, this seems to kick the DNS server and it will work again for a while.

This has been a problem for the last year+, and we have tried to debug and figure out what could be causing it with no luck, it happens only for some of us, and we have same laptops, software, connections, etc
We have also transitioned to new laptops 6 months ago, and still the same pattern persists, only some experience it consistently, while others never see it, and it happens both in the office and at home for people.

We are all running latest Windows 11, connected to the same 1-3 PFSense OpenVPN servers.
Is anyone else seeing this type of issue? I tried to reach out to your support email late last year with no luck, so trying here now as the issue is becoming increasingly annoying for everyday work for those affected.
It does not happen to me personally, but I can ask the team to collect what will help with this issue.

Kind regards
Niels

Hi,

If you’ve already emailed us in the past then it’s likely we’ve already recommended similar troubleshooting steps to those outlined below, however they do cover the possible causes so please do make sure you run through them with any end users experiencing problems.

Firstly, make sure that they’re using the latest version of Viscosity (1.11.5 at the time of writing). If they’re using an old version they may be running into bugs or compatibility issues. Also make sure their copy of Windows is up to date with the latest updates.

You’ll then want to make sure they’re actually experiencing a DNS related issue, rather than a different network issue. When trying to access something like a webpage, the first thing the computer does is attempt to perform a DNS lookup: if there is a different network problem (such as a routing issue) then this lookup will likely fail and the browser may display a DNS/lookup timeout error message even though the actual underlying problem is something else.

When a user finds that their VPN connection is no longer working they should try running through the “Checking For A DNS Problem” and “Checking For A Routing Problem” sections in the following article to see whether it could be a network related issue. For example, the underlying VPN connection may be dropping out, or they may have a routing conflict when a DHCP renew on their local network takes place, etc.
https://www.sparklabs.com/support/kb/article/troubleshooting-connection-problems/#unable-to-use-the-vpn-connection

Also keep in mind that OpenVPN may not detect a dropped VPN connection immediately (it depends on the ping/ping-restart values), so when a user’s network connectivity is lost have them try waiting several minutes to see whether OpenVPN detects that a drop-out has occurred in the log (typically with more information).

If after running through the above steps you find that it does appear to be a DNS issue, then the following article covers the possible causes. You’ll only want to follow the troubleshooting steps listed in the article while the VPN connection isn’t working, as you likely won’t gain much insight if the VPN connection is in a working state.
https://www.sparklabs.com/support/kb/article/troubleshooting-dns-problems/

As the above article also touches on, you’ll want to check for any potentially conflicting software. This is particularly important as you mention only some users are experiencing issues, so it may be due to something they have installed on their machine. Such software can often react to network changes, or periodically apply settings, which could be causing network or DNS issues.

In particular check for installed Endpoint Security Software. Such software includes things like Antivirus software, firewall software, filtering software, device management software, DNS proxies, certain networking tools and drivers, and some other VPN clients. Some endpoint security software (typically installed on enterprise machines) can be quite aggressive and block or inject itself into system processes, which can cause issues. If any such software is installed, try temporarily disabling it or whitelisting/excluding Viscosity and see whether the issue persists. You may need to uninstall/reinstall Viscosity after uninstalling/disabling any such software.

It’s also worth checking the Windows Event Viewer for any errors or warnings around the time a user loses network connectivity, not just from Viscosity but from Windows and all applications.
https://www.sparklabs.com/support/kb/article/viewing-the-event-viewer/

Finally, don’t forget to check the logs on your OpenVPN server (as well as DNS server etc.). Look for log messages around the time a user reports lost network access. It’s possible that this may be unrelated to the end user’s computer, and something on the server is changing and causing the network/DNS issues (such as conflicting connections using the same credentials).

Regards,
Aaron

Thank you very much Aaron,

I am confident there is a DNS issue, when running “lookup google.com” it will show viscosity and timeout.
We are not using any 3rd party antivirus or similar, and all installations and OSs are up-to-date.
We mostly have the same software installed, and things like other vpn clients or proxies are not used afaik.

We also route only a single subnet usually through each VPN connection to cloud resources, but all other traffic should go through normal connection to the internet.

I will try and get some to monitor the logs (app and windows) for when it happens again and check out the debugging steps in the articles.

Really appreciate it :slight_smile:
Kind regards
Niels