Important note about SSL VPN compatibility for 20.0 MR1 with EoL SFOS versions and UTM9 OS. Learn more in the release notes.

This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Sophos Connect v2.3 and SFOS v20 MR1 - SSL VPN - Delayed disconnection

Hi,

With the noon version of Sophos Connect v2.3 against SFOS v20 MR1, I encountered a bug in the delayed disconnection of SSL VPN server-side (XG Sophos firewall). On the client side the connection is already in a disconnected state, but on the firewall I can still see an active connection. This state lasts about 5 min. I compared this with the behavior with OpenVPN community client v2.6.10 and there the disconnection occurs on the server side a few seconds after the client side.



This thread was automatically locked due to age.
  • Hi,  Thank you for reaching out to the Sophos community team. I have tried to check a similar kind of re-production with SF OS and Connect client version details mentioned by you but unfortunately, the issue is not getting re-produced for me in my setup.

    i.e. If I click on "Disconnect" on the Sophos Connect client side, the same user is removed from live users immediately. 

    If the client is disconnected due to "Dead peer detection" then the firewall will close the connection after 180 seconds (3 min - default time) with an unresponsive client or whatever time has been set by the admin in  SSL VPN global settings.



    For more info: Sophos Firewall: Understanding the Idle timeout and the dead peer detection for remote access SSL VPN
    support.sophos.com/.../KB-000038126

    Regards,

    Vishal Ranpariya
    Technical Account Manager | Sophos Technical Support

    Sophos Support Videos | Knowledge Base  |  @SophosSupport | Sign up for SMS Alerts |
    If a post solves your question use the 'Verify Answer' link.

  • Hi 

    Thanks for the response. That's strange that it can't be reproduced, because I've had it behave that way on three PCs for two users so far. Maybe I can enable the SSL VPN debug mode and send you the output from it.  

  • Hi   

    I still don't want to terminate the OpenVPN connection on the server side when using the Sophos Connect client. OpenVPN Connect v3.4.4 and OpenVPN commnutity client v2.6.10 does not behave this way and the server side connection is terminated within seconds on the client side.

    Should I open a standard support ticket for this?

  • Hi  Thank you for this latest update, yes please open a support case, so this can be checked and validated further and please share the case ID with me via DM or here, so I can overlook the progress of the case and if any internal notes are needed from my end I may add over the case, thanks.

    Regards,

    Vishal Ranpariya
    Technical Account Manager | Sophos Technical Support

    Sophos Support Videos | Knowledge Base  |  @SophosSupport | Sign up for SMS Alerts |
    If a post solves your question use the 'Verify Answer' link.

  • Hi  

    I now believe that this is error NC-120615. However, I have not encountered this error before because I am using the OpenVPN community client which does not have this problem. Only I'm not allowed to use the reconnect function with it, which doesn't close the original connection each time, but creates a new one. Other users using Sophos Connect mi have not encountered this error, as they usually do not make multiple connections/disconnections in a short period of time when the maximum number of simultaneous connections could be exhausted. Globally we have 3 simultaneous connections set up.  

  • Hi  Thank you for sharing this information, if your existing SSL VPN is configured on UDP protocol then yes there may be a chance it is related to NC-120615.

    Currently, Sophos connects clients with UDP-based SSL VPN configuration, with no feature there to send the SIGTERM to the server while for TCP-based connection, TCP RESET packets notify the server for the same.

    If you are using UDP-based SSLVPN RA then the workaround would be to set a TCP-based SSLVPN RA connection, which will immediately disconnect the user from the live user list.

    In my LAB SSL VPN is set to TCP and that is the reason the issue was not observed for me when I tried to check/re-produced.

    Regards,

    Vishal Ranpariya
    Technical Account Manager | Sophos Technical Support

    Sophos Support Videos | Knowledge Base  |  @SophosSupport | Sign up for SMS Alerts |
    If a post solves your question use the 'Verify Answer' link.

  • Hi  

    Yes, we intentionally use UDP for SSL VPN because we also transmit video streams (usually in unicast UDP) from PTZ (Pan - Tilt - Zoom) cameras over VPN, where UDP protocol has lower latency.

    I have verified on a test appliance (XG210) that after switching the VPN to TCP, the problem no longer occurs with the Sophos Connect client.

  • Hi  Thank you for sharing the details of the quick test results on the test appliance. Appreciate your quick work on this for validation.

    Regards,

    Vishal Ranpariya
    Technical Account Manager | Sophos Technical Support

    Sophos Support Videos | Knowledge Base  |  @SophosSupport | Sign up for SMS Alerts |
    If a post solves your question use the 'Verify Answer' link.