Inconsistent IPv6 gateway behavior after prefix change
-
Hi,
I'm running into a pretty niche IPv6 issue and could use some insight. Not sure if it's worth opening a thread, but here we go.
Setup overview:
-
MikroTik router handling PPPoE
-
pfSense sits between MikroTik and LAN clients
-
Both ISPs provide dynamic /64 IPv6 PDs (not very useful), so I use NAT66 with some scripting to update NAT rules when prefixes change
Behavior on pfSense 2.7.2: Everything works great — load balancing, failover, PBR — all smooth across both ISPs with IPv6.
But on pfSense 2.8.0, I'm seeing some strange behavior when ISP1's prefix changes:
- Case 1 (Problem):
-
ISP1's prefix changes
-
pfSense interface gets new SLAAC address
-
My scripts correctly update NAT66 rules
-
However, the IPv6 gateway remains marked as offline until I manually restart it from System > Routing (restart dpinger)
- Case 2 (Works Fine):
-
ISP2's prefix changes
-
SLAAC + NAT66 updates go through cleanly
-
Gateway comes back online automatically
- Case 3 (Weird Workaround):
-
If ISP1’s gateway is offline after a prefix change
-
I manually force a DHCPv6 release/renew on ISP2 (not ISP1) from MikroTik
-
Both gateways (ISP1 + ISP2) come back online in pfSense
Also worth noting: in MikroTik's IPv6 neighbors list, most of ISP1’s entries appear as "unreachable" or "stale" when this issue occurs.
This issue is only present on pfSense 2.8.0. Everything behaves as expected on 2.7.2 with the same exact config and scripts.
After discussing this with @stephenw10, we realized it’s due to /etc/rc.newwanipv6 not firing and therefore not restarting dpinger.
So right now, I have to manually run that every time the IPv6 prefix changes — which, thanks to my amazing ISP, happens every 24 hours.
-
-
@milindhvijay Not fixed in the latest beta.