Project

General

Profile

Actions

Bug #7156

closed

Change in 'Block bogon networks' or 'Block private netowrks' GUI options kills routing entries for OpenVPN interfaces.

Added by Karl Fife over 7 years ago. Updated about 7 years ago.

Status:
Rejected
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
Start date:
01/23/2017
Due date:
% Done:

0%

Estimated time:
Plus Target Version:
Release Notes:
Affected Version:
All
Affected Architecture:
All

Description

It appears that toggling in the 'Block bogon networks' and/or 'Block private netowrks' GUI option kills the automatic routes inserted for openvpn server service (and/or client service).

Steps to recreate (server-side example):
- Validate that server side routes exist for OpenVPN server assigned to interface ovpns1
e.g. 10.1.128.0/17 172.30.1.130 UGS 40038923 1500 ovpns10
- ENABLE 'Block bogon networks' on the corresponding interface. Save, Apply.
- Refresh server-side routes.
- Note the route is absent, and that new connections fail.
- Now RESTART the OpenVPN server service.
- Note that the routes have been re-inserted into routing table, and new connections can again be established.

I observe the same behavior again when the 'Block bogon networks' option is DISABLED
I observe the same behavior when toggling 'Block private networks' instead of 'Block bogon networks'
I observe the same behavior when toggling client side interfaces for OpenVPN client service.
I have observed this on multiple versions including 2.2.6 (amd64-Full) and 2.3.2-RELEASE-p1 (i386-nanobsd).

Actions #1

Updated by Jim Pingle over 7 years ago

  • Status changed from New to Rejected

Any time you save/apply changes on an assigned OpenVPN interfaces you have to restart the VPN. It's always been that way. Nothing to do with the options you select.

Actions #2

Updated by Karl Fife over 7 years ago

How is one supposed to know that this is a requirement? It's exactly not self-evident that changing something as simple as the interface name would break OpenVPN functionality (which in fact, it appears to do). I've never seen it documented.

Shouldn't that seemingly idiosyncratic detail be documented somewhere to prevent prevent disruptive lockouts? I haven't ever read about this requirement. I think a case could be to have the 'save' action conditionally restart the OpenVPN service, if OpenVPN is the underlying interface type (e.g. ovpns{n}).

Actions #3

Updated by Jim Pingle over 7 years ago

It could probably be documented in a way that calls more attention to it, but it is mentioned in the documentation where assigning OpenVPN interfaces is covered: https://portal.pfsense.org/docs/book/openvpn/assigning-openvpn-interfaces.html#interface-assignment-and-configuration

Actions #4

Updated by Karl Fife about 7 years ago

As I've thought about this more, it seems to me that the correct behavior is for "Apply Changes" to trigger a restart the OpenVPN service if changes are made to the interface, and if the underlying interface type is OpenVPN.

The case to be made is that otherwise it's not possible to make a changes to a remote interface without a lockout condition. Documentation/knowledge of the restart requirement is irrelevant.

Even if it were possible for a remote administrator to restart OpenVPN, it seems more consistent with the goal of the project in general to have the GUI to restart/reload any underlying dependencies/rules/services when changes are applied, rather than relying on a sysadmin to keep track of such things.

Actions

Also available in: Atom PDF