Closed pmelange closed 5 years ago
This issue might have to be reported against the external-repo we use.
the origional ip rule
config rule
option lookup 'main'
option priority '200'
option in 'dhcp'
doesn't work anymore. But this rule does (unfortunately with hardcoded ip range)
config rule
option lookup 'main'
option priority '200'
option dest '192.168.0.0/24'
Now that a solution has been found, closing....
Hedy 1.0.2 tunneldigger version.
Having a management network running parallel and on the same VLAN as another network results in not being able to route to the second network space. For example, quitte-core has a normal 10.36.x.x/24 client network. Additionally there are many AP's on the network, and all have been given an ip in the 192.168.0.0/24 range.
Previously, the firmware was able to route this properly. Here is what the rule looks like now.
It should look like
I have confirmed with tcpdump that outgoing ping requests are being sent to 192.168.0.21 with a source address of 10.36.x.1.
192.168.0.x/24 shows up in the following tables
Quitte-core also has a second setup for a mesh device. First, the OLSR mesh address, and second a management address in the 192.168.3.x/24 range. Same problem.
I have tried setting this up in multiple ways. I have tried aliases
option ifname '@dhcp'
, explicit device namesoption ifname 'eth0.10'````, setting up a bridge, and having 2 IP's in the dhcp interface config
list ipaddr 'xxxx'```.The workaround is to remove any interfaces which share a device (the second network) from the configuration. Then, manually add the IP address after boot.
The previous firmware image was a hand made ipsec image created by @SvenRoederer with a version number of 1.0.0 alpha. Having multiple address spaces on one device worked fine. I believe that since then, the policy routing has been changed quite a bit.