Port 443 timeout using Netcat but is working in browser

[Frogg]

Sorry i didnt told it, a simple nat redirect
WAN * 443 to LAN localip 443

With equivalent rule in firewall

I will updated main post with screen shots after post restriction

Firewall

Nat

[johnpoz]

So nat reflection, what type pure proxy? If you do not do the correct nat reflection you run into asymmetrical problems..

Does it work from outside? Test with can you see me . org

Is pfsense listening on 443? It normally listens on 80, etc..

Why would you not just resolve domain.tld to your local IP via your local dns, host override vs doing a reflection, at best your hairpinning all your traffic for no reason. And can run into issues with asymmetrical traffic flow.

[Frogg]

Thanks for your reply

PFSense web interface has been changed to another port (https 8080)

All 443/80 trafic is redirected to 192.168.1.4 which is an Apache displaying content from domain name sent in the client request.

In fact all is working well (or it seems to works) website are displaying on the browsers.

The trouble come from tools like netcat, telnet, openssl connect, etc ... on port 443 going timeout like if PFSense was bloking the response

Same trouble with all other ports. Only port 80 seems to respond

For exemple:

nc -v externalIP 111
nc: connect to externalIP port 111 (tcp) failed: Connection timed out

but FTP Client works well (NAT 111 WAN to 21 LAN)

nc -v domain.com 80
Connection to domain.com 80 port [tcp/http] succeeded!

The only port working as attempt

[johnpoz]

Did you turn off the 80 redirect on pfsense? pfsense will attempt to redirect traffic to "it" on 80 to its ssl port.

If browsers and clients working - then what is the problem?

[Frogg]

What do you mean by turn off 80 redirect ?
There is the NAT on *:80 WAN to 192.168.1.4:80 LAN with firewall rules for it.
The PFSense Web site is on port 8080 https.

The trouble come from network supervision like Zabbix:

2:58:54  High   PROBLEM   Service web https is down on 
12:58:52 High   PROBLEM   Service web https is down on 	
12:58:50 High   PROBLEM   Service web https is down on 
....

And other test cant be done like SSL days left on validity for exemple

With classical NAT (physical router) this same schema is working. I really fell like PFSense is blocking the response for a reason i don't know

[johnpoz]

It wouldn't be "blocking" do a sniff.. Clearly there is a response since you say browsers and clients work.

But they might be more forgiving if response comes from different mac - asymmetrical.

[Frogg]

Thanks,

i missed this one, now all port doesn't not respond (but it is ok)

I just saw something:
Netcat is working as it should for external network, and all is ok

ExcAst:~ # nc -v domain.com 443
Connection to domain.com 443 port [tcp/https] succeeded!

ExcAst:~ # nc -v domain.com 80
Connection to domain.com 80 port [tcp/http] succeeded!

ExcAst:~ # nc -v domain.com 777
Connection to domain.com 777 port [tcp/multiling-http] succeeded!
SSH-2.0-OpenSSH_7.2p2 Ubuntu-4ubuntu2.8

The problem come from internal netcat request (LAN)

maybe it is link to this:

DNS fwd/rev mismatch: domain.com != machinedomain

Websites are working in Lan browsers

[johnpoz]

@Frogg said in Port 443 timeout using Netcat but is working in browser:

DNS fwd/rev mismatch: domain.com != machinedomain

which is why what your doing is not the correct solution.. What you should be doing is dns resolution to access your machines via their local IPs vs nat reflection..

[Frogg]

Do you have any clue on how i could achieve something like that ?

[johnpoz]

Yeah a simple host override so that whatever.domain.tld resolves to the actual local IP you want 192.168.1.4 I take it.

[Frogg]

I added the domain name to the local 192.168.1.4 machine if i check /etc/hostname i get the domain name

But i don't think it would fix it has the domain name in the comparaison il the host xx.xx.xx.xx.rev.poneytelecom.eu reverse DNS

Or i am really missing something

[johnpoz]

On pfsense create a host override - your clients point to pfsense for dns right?

So whatever.domain.tld points to your local IP 192.168.1.4, and then the ptr for 192.168.1.4 would also resolve to that..

[Frogg]

Thanks for the help,
but i fill like i am lacking of knowledge on this part.

In network config of 192.168.1.4 (and all local machines) i use 192.168.1.1 as DNS

In the interface DNS resolver was disabled, so i enabled it and added the Host Overrides

Dig domain.com

;; ANSWER SECTION:
domain.com      960     IN      A       internetIp

Dig - x domain.com

;; AUTHORITY SECTION:
in-addr.arpa.           3600    IN      SOA     b.in-addr-servers.arpa. nstld.iana.org. 2020081164 1800 900 604800 3600

i think i am missing something to have the same result as yours

[johnpoz]

@Frogg said in Port 443 timeout using Netcat but is working in browser:

i use 192.168.1.1 as DNS

And what is that? Is that not pfsense? This only works if what your using for dns for your local network actually ends up asking pfsense.. So the host override can be returned.

If your network is using 192.168.1.1 for dns, then setup your dns records there.

[Frogg]

Yes 192.168.1.1 is PFSense
Maybe it require a reboot, i ll give a try

[johnpoz]

You said the resolver was disabled - where you using forwarder (dnsmasq) If so then the host override would be setup there and not in the resolver (unbound). Both forwarder and resolver allow for host overrides. Set the override in whichever one your using.

You only can have either or running forwarder or resolver. You can not run both listening on 53.

A reboot is not required that is for sure.. But you may have issue with local caching on the client, so you would have to flush the local dns cache of whatever client your testing from.

[Frogg]

Thanks again for the help!

Forwarder & Resolver (Now changed to enabled) was disabled

It seems i am using the default cache (so it should be flushed with reboot)

systemd-resolve --flush-caches
Failed to flush caches: Unit dbus-org.freedesktop.resolve1.service not found.

After reboot result is the same

any idea ?

[johnpoz]

@Frogg said in Port 443 timeout using Netcat but is working in browser:

Forwarder & Resolver (Now changed to enabled) was disabled

Huh? You can not use both at the same time.. You run into a race condition.. Which one are you using? Place your host override in the one your using.. They both allow for overrides.

Do a directed query to pfsense to validate it returns your records you put in host override..

C:\>dig @192.168.9.253 ahost.domain.tld

; <<>> DiG 9.16.6 <<>> @192.168.9.253 ahost.domain.tld
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 8719
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;ahost.domain.tld.              IN      A

;; ANSWER SECTION:
ahost.domain.tld.       3600    IN      A       192.168.1.4

;; Query time: 0 msec
;; SERVER: 192.168.9.253#53(192.168.9.253)
;; WHEN: Tue Sep 08 13:40:59 Central Daylight Time 2020
;; MSG SIZE  rcvd: 61


C:\>nslookup
Default Server:  pi-hole.local.lan
Address:  192.168.3.10

> server 192.168.9.253
Default Server:  sg4860.local.lan
Address:  192.168.9.253

> ahost.domain.tld
Server:  sg4860.local.lan
Address:  192.168.9.253

Name:    ahost.domain.tld
Address:  192.168.1.4

Pfsense in my case is 192.168.9.253