Problem with snapshot 3/31 16:19:49

JonathanLee

This post is deleted!

jimp

Please post the text and not screenshots. It's making it unnecessarily hard to follow.

Looking at the code it looks like it's from NAT reflection not outbound NAT. Is that a port forward for NTP you have? Is NAT reflection enabled globally or perhaps on just that rule?

JonathanLee

@jimp You got it thanks for the help

here is your request this was the original settings, I have tried to delete the one line however it re adds

Outbound NAT rules (automatic)

Subnets to NAT

tonatsubnets = "{ 127.0.0.0/8 ::1/128 192.168.0.0/16 }"
nat on $WAN inet from $tonatsubnets to any port 500 -> 64.113.111.129/32 static-port
nat on $WAN inet6 from $tonatsubnets to any port 500 -> (mvneta0) static-port
nat on $WAN inet from $tonatsubnets to any -> 64.113.111.129/32 port 1024:65535
nat on $WAN inet6 from $tonatsubnets to any -> (mvneta0) port 1024:65535

TFTP proxy

rdr-anchor "tftp-proxy/*"

NAT Inbound Redirects

rdr on mvneta1 inet proto { tcp udp } from 192.168.0.0/16 to any port $DNS -> 192.168.1.1
no nat on mvneta1 inet proto tcp from (mvneta1) to 192.168.0.0/16
nat on mvneta1 inet proto tcp from 192.168.0.0/16 to 192.168.1.1 port $DNS -> (mvneta1)

rdr on mvneta1 inet6 proto { tcp udp } from any to any port $DNS -> ::1
rdr on mvneta1 inet proto udp from 192.168.0.0/16 to any port 123 -> 192.168.1.1
no nat on mvneta1 inet proto tcp from (mvneta1) to 192.168.0.0/16
nat on mvneta1 inet proto tcp from 192.168.0.0/16 to 192.168.1.1 port 123 -> (mvneta1)

rdr on mvneta1 inet6 proto udp from any to any port 123 -> ::1

Setup Squid proxy redirect

rdr pass on mvneta1 inet proto tcp from any to !(mvneta1) port 80 -> 127.0.0.1 port 3128
rdr pass on mvneta1 inet proto tcp from any to !(mvneta1) port 443 -> 127.0.0.1 port 3129

JonathanLee

@jimp

This is my NAT for DNS and NTP

I force them to use the firewall

This setting has worked for some time now I get time from the firewall no matter what wan ip address a device requests for NTP. I had issues with time being jumped 15-20 mins without this nat setting for NTP, no clue why so I added a rule that let the firewall handle all NTP requests, after that no issues with time jumps

jimp

Do you have NAT reflection enabled globablly or on those rules? If so, disable it.

JonathanLee

@jimp

Network Address Translation Settings for Firewall

All is disabled

jimp

There is also a per-rule option on the individual port forwards. Is it enabled there?

jimp

Also your port forward target should really be 127.0.0.1 and not 192.168.1.1, assuming 192.168.1.1 is the LAN address of the firewall.

JonathanLee

@jimp

Firewall rules

JonathanLee

@jimp said in Problem with snapshot 3/31 16:19:49:

127.0.0.1

correct to loopback address

filter reload failed at same point as before with loopback address

jimp

Did you reboot after making that change? Once you get the error you have to reboot to know if it's fixed. Also check and see if those no nat on rules are still present.

JonathanLee

@jimp

I have to wait for 2 hours to do a reboot I am at work and the firewall must be up for 2 hours. I have changed the NTP and DNS to use the loopback 127, the no nat rules still show in debug if you delete them they return. Thanks I will reboot at 11:10am

jimp

OK, I see the real source of your duplicate rules now and I can replicate it here. Your destination on those port forwards is Any and it should not be set to that. You should set that to ! LAN address (so check the Invert Match box for the destination and then set the destination to LAN Address)

As described in the docs:
https://docs.netgate.com/pfsense/en/latest/recipes/dns-redirect.html

JonathanLee

@jimp

I have corrected this to inverted

jimp

That looks good, if you look in /tmp/rules.debug now you shouldn't see those no nat on rules under the rdr rules for those port fowards. If that is the case, you should be in good shape after a reboot.

JonathanLee

@jimp

They are gone now, I wonder why that worked for so long before this update.

Outbound NAT rules (automatic)

Subnets to NAT

tonatsubnets = "{ 127.0.0.0/8 ::1/128 192.168.0.0/16 }"
nat on $WAN inet from $tonatsubnets to any port 500 -> 64.113.111.129/32 static-port
nat on $WAN inet6 from $tonatsubnets to any port 500 -> (mvneta0) static-port
nat on $WAN inet from $tonatsubnets to any -> 64.113.111.129/32 port 1024:65535
nat on $WAN inet6 from $tonatsubnets to any -> (mvneta0) port 1024:65535

TFTP proxy

rdr-anchor "tftp-proxy/*"

NAT Inbound Redirects

rdr on mvneta1 inet proto { tcp udp } from 192.168.0.0/16 to !192.168.1.1 port $DNS -> 127.0.0.1
rdr on mvneta1 inet proto udp from 192.168.0.0/16 to !192.168.1.1 port 123 -> 127.0.0.1

Setup Squid proxy redirect

rdr pass on mvneta1 inet proto tcp from any to !(mvneta1) port 80 -> 127.0.0.1 port 3128
rdr pass on mvneta1 inet proto tcp from any to !(mvneta1) port 443 -> 127.0.0.1 port 3129

UPnPd rdr anchor

rdr-anchor "miniupnpd"

JonathanLee

@jimp

Thank you for your help with this and all you do.

jimp

The update has a problem with duplicate rules. We're still working on a fix, but in the meantime you can work around it by eliminating the duplicate rules. Most people get them from reflection + VIPs, but in your case it was port forwards with a destination of Any.

I'm surprised those port forwards worked well at all, they would have been looping traffic around a bit since even a query sent to the firewall would be caught and then redirected to ... the firewall. Unnecessarily. Port forwards for a destination of Any have a long history of problems.