Firewall rule not blocking
I’m at my wit’n end trying to trouble shoot this issue:
Attached is the firewall rule for the lab_integration interface (10.1.0.1./16)
Configuration is as follows:
1. Multiple VLANs assigned to the physical LAN device.
2. Each VLAN on a separate and unique subnet
3. Single WAN interface
Attached screen shot firewall rule description:
1. Allow ICMP to any interface
2. Allow DNS lookup to the internal DNS servers on a different subnet. The alias AMGT_DNS defines two different DNS servers on the 10.0.0.0/16 subnet
3. Allow NTP sync to the internal NTP server on a different subnet. The alias NTP_SERVER points to 10.0.1.20.
4. Allow NFS interface to the internal NFS server on a different subnet. The alias Internal_Servers defines two different NFS servers on the 10.0.0.0/16 subnet
5. Allow SSH/SCP interface to the Users address (10.0.0.0/16).
6. Block everything else.
Based on this set of rules, I would expect the firewall to not allow any traffic to the Internet. However, when I run “curl www.google.com”, I get the response back from google. I looked at the firewall logs and nothing show us. However, when I look at the diagnostics–>states and filter on the machine on the 10.1.0.0/16 network (machine IP address: 10.1.200.1), I see that NAT occurred from 10.1.200.1–>127.0.0.1:xxxx . I don’t get this at all.
I could really use the help figuring this out.
![Screen Shot 2017-12-06 at 10.16.06 PM.png_thumb](/public/imported_attachments/1/Screen Shot 2017-12-06 at 10.16.06 PM.png_thumb)
![Screen Shot 2017-12-06 at 10.16.06 PM.png](/public/imported_attachments/1/Screen Shot 2017-12-06 at 10.16.06 PM.png)
You don’t need that last rule since there is a hidden Default Deny rule on all interfaces.
Did you reset your states after you made your firewall rule changes? Established states will not be affected by a rule update.
Yes. I did reset the states after the firewall rule changes. No changes in behavior.
I’m scheduling a reboot of the pfsense box this weekend to see if that clears it up.
Those rules should block WWW traffic on that interface. Are you sure you’re on the interface and not some other VLAN?
Yes. I’m sure. Just for yucks, I moved the “block all” rule to the top and it stopped the curl command.
I will be rebooting the pfsense box tonight and see if that clears things up.
In addition, from the machine on the VLAN, here’s the default routes:
[root@localhost ~]# route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
10.1.0.0 0.0.0.0 255.255.0.0 U 1 0 0 eth0
0.0.0.0 10.1.1.1 0.0.0.0 UG 0 0 0 eth0
Also, when I ran “curl www.google.com”, here’s the firewall log and the state log:
2 Matched Firewall Log Entries. (Maximum 2000) Pause
Action Time Interface Source Destination Protocol
Dec 7 10:29:42 LAB_INTEGRATION 10.1.200.1:42270 10.0.1.22:53 UDP
Dec 7 10:29:08 LAB_INTEGRATION 10.1.200.1:45254 10.0.1.22:53 UDP
Interface Protocol Source (Original Source) -> Destination (Original Destination) State Packets Bytes
LAB_INTEGRATION udp 10.1.200.1:58220 -> 10.0.1.22:53 SINGLE:MULTIPLE 2 / 2 120 B / 164 B
USERS udp 10.1.200.1:58220 -> 10.0.1.22:53 MULTIPLE:SINGLE 2 / 2 120 B / 164 B
LAB_INTEGRATION tcp 10.1.200.1:55985 -> 127.0.0.1:3128 (188.8.131.52:80) FIN_WAIT_2:FIN_WAIT_2 16 / 15 1017 B / 13 KiB
LAB_INTEGRATION tcp 10.1.200.1:813 -> 10.0.1.11:2049 ESTABLISHED:ESTABLISHED 28.851 K / 24.415 K 5.11 MiB / 5.83 MiB
USERS tcp 10.1.200.1:813 -> 10.0.1.11:2049 ESTABLISHED:ESTABLISHED 28.851 K / 24.415 K 5.11 MiB / 5.83 MiB
USERS tcp 10.0.4.1:52516 -> 10.1.200.1:22 ESTABLISHED:ESTABLISHED 359 / 233 26 KiB / 98 KiB
LAB_INTEGRATION tcp 10.0.4.1:52516 -> 10.1.200.1:22 ESTABLISHED:ESTABLISHED 359 / 233 26 KiB / 98 KiB
One of my main question is the route to 127.0.0.1 on the third line of the filtered states log.
That’s a redirect to squid web proxy which listens on tcp/3128.
That was the key. What I thought was seeing was the cached page for the sites from the squid proxy.
Thanks for getting my head out of my #@$.
Glad to help.