Project

General

Profile

Actions

Bug #7326

closed

Unbound fails to start during rc.wanipchange when using large enough dns lists

Added by Chris Collins over 4 years ago. Updated over 4 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
DNS Resolver
Target version:
Start date:
02/27/2017
Due date:
% Done:

100%

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

Description

Current pfsense code manually kills unbound, waits just one second and then starts it again, however if unbound doesnt shut down quick enough following the kill command it will fail to start with an error it cannot bind the port because it is already in use. Stopping unbound with a kill command is not good practice anyway and should be a last resort, by using the unbound-control stop command instead, pfsense will actually wait for the shutdown process to complete before starting unbound again and it will succeed. I have patched my local system and can share the patch here, but I am not a php developer so my changes may not be in the best syntax and for some reason I can see from the logs when the rc.newwanip script runs it tries to shutdown unbound 3 times quickly so the log will show it failing twice (first time works), to keep things tidy its probably also a good idea to look at what is triggering the unbound reconfigure code 3 times.

With this bug (undesirable behaviour), the following situations occur.

1 - On bootup unbound will be offline as when the wan comes online it will trigger the rc.newwanip script and it will fail (without my patch).
2 - If WAN is cycled, unbound will also be left in a offline state.


Files

unbound.patch (739 Bytes) unbound.patch Chris Collins, 02/27/2017 11:23 AM
unbound.patch (744 Bytes) unbound.patch Chris Collins, 02/27/2017 11:27 AM
unbound-stop.diff (1.47 KB) unbound-stop.diff Jim Pingle, 07/07/2017 01:42 PM
Actions

Also available in: Atom PDF