OPNsense Forum
Archive => 18.1 Legacy Series => Topic started by: bladman on April 19, 2018, 12:54:29 pm
-
Hi,
We plan on taking a new fiber connection into production.
So far so good, but currently I am testing the new connection and I am experiencing some issues.
I have set the outbound NAT to Hybrid. I have set one machine to be routed over the new connection with as address an IP Alias, but it appears I still am being routed over the current primary WAN interface.
When I set a rule for this specific machine with the gateway towards the new fiber I experience a lot of issues with visiting websites where it seems that I have some sort of split routing (some pages load partially but most pages don't load at all).
A traceroute to the IP Alias I have setup on this machine shows the traffic going through all the proper hops but after one of the hops the (current) main IP address of the WAN IP appears, the traffic is forwarded to the default gateway of the current WAN interface and then being routed again. So a routing loop seems to appear.
Any ideas?
-
Have you tried unsetting "Use sticky connections" in Firewall->Advanced as a test?
This has caused us grief of similar sorts in MultiWAN setups.
-
Thanks for your reply but unfortunately this allready has been unset.
-
Some more information, the load balancer virtual server statuses show Unknown - relayd not running? as well.
Seems to be related.
-
No, outbound load balancing does not use relayd, but route-to from pf.
Can you post a screenshot of your outbound GW redirect rule?
-
Hi,
We plan on taking a new fiber connection into production.
So far so good, but currently I am testing the new connection and I am experiencing some issues.
I have set the outbound NAT to Hybrid. I have set one machine to be routed over the new connection with as address an IP Alias, but it appears I still am being routed over the current primary WAN interface.
When I set a rule for this specific machine with the gateway towards the new fiber I experience a lot of issues with visiting websites where it seems that I have some sort of split routing (some pages load partially but most pages don't load at all).
A traceroute to the IP Alias I have setup on this machine shows the traffic going through all the proper hops but after one of the hops the (current) main IP address of the WAN IP appears, the traffic is forwarded to the default gateway of the current WAN interface and then being routed again. So a routing loop seems to appear.
Any ideas?
Did you set both WANs as the same tier?
Cheers Robert
-
Hi,
The loadbalancing is inbound, not outbound.
Behind the loadbalancer are for example two Exchange servers providing Client Access.
On the current "main" connection everything works just fine and relayd reports those VS's as up.