I’d wager that you would be in with a good chance. 12000 is obvs an example.
Just remember to tell safenode which port to use. I have a suspicion that folk are omitting that part.
2 Likes
I don’t like the limitation on setting the protocol and the fact the DMZ wasn’t working makes me want to isolate the machine running the node software by giving it a direct internet connection.
1 Like
I’ve tried connecting directly to the modem/router, rather than my other router, and unfortunately it’s the same with that. With port forwarding set up, and DMZ, and Windows Firewall off, no ports are visible externally, so I have to assume it’s the ISP messing things up, but I will keep looking for solutions!
One interesting thing I don’t understand due to my ignorance, is that the modem/router has a different external IP address to the other router that’s connected via Bridge mode to the modem/router. I guess that’s normal, but I expected one internet connection would have one external IP, not 2.
I’d ask what the ip addresses are but then it sounds like i’m phishing. I’m just wondering if they are both routable addresses. Maybe a less intrusive question, ISP? IPv6, IPv4, both? They aren’t on a weird internal network and nothing along the lines of 10.0.x.x, 172.16.x.x, or 192.168.x.x ?
1 Like
The ISP is MEO in Portugal, and the external IPs look like normal external IPv4 IPs for an internet connection, definitely not internal IPs like those you noted.
Not to worry - I’ll keep trying, but if no luck, I’ll just need to use a VPS for nodes… or wait for NAT traversal to be ready.
2 Likes
The one your PC is looking for.
That is normal for ISPs. It typically has the same behaviour as your home router with DHCP. The device is allocated an IP address for a certain period. The device’s MAC address is used to identify the device and if the device is not seen for that period of time then the allocation is forgotten.
In bridged mode then your PC is supplying the MAC address used for allocation and thus gets its own IP address which typically will be different to the Modem’s allocated IP address since its allocation hasn’t timed out yet at the ISP end.
1 Like
Finally reset my router to setup the forwarding and actually try to join a test network for once. hooray for me!
3 Likes
I can set protocol to all and hopefully there shall be no issues.
Now make sure your firewall allows both.
2 Likes
So, just the same for internal and external in this case? E.g. 12000.
Unfortunately it makes no difference for me & port forwarding can’t seem to make anything visible externally (assuming there isn’t a problem with the port forwarding tester).
1 Like
Not working in the last testnet either. So since the switch to UDP:
And this is my router config:
The last 3 entries correspond to the raspberry pi. If you check the two entries not related to SAFE, there’s one for TCP working (SSH) and one for UDP working (openvpn). So yes, my port-forwarding is working, because I’m using it, and in both protocols.
The two other entries (the ones on top) are from other machines in my LAN. The port range is the same than the one I’m using in the script from @aatonnomicc (so I entered XY700 as the initial port, the same as in the range shown there). For having things easier, the range is the same internally and externally.
Any ideas here?
And BTW, this very same setup was working for TCP.
1 Like
just a thought I see you have open vpn working on udp
change the ports from BOTH to UDP only in case there is something up with your routers implementation of both
also can you run sudo ufw status on your system and post the out put
just incase the scrip has not added the fire wall rule correctly. with the swithc from tcp to udp perhaps the script never deleted the old rules for tcp so can not add a new rule on the same port for udp.
It was stopped, however I had stopped the borders and recreated them when I checked, because I only read your previous message. When I started it, I could see the right rule. As I restarted minutes ago, I’m going to restart everything from the script again. If I get patient enough, I might recreate the container…
Thanks!
Edit: I see the rule, and the ports on it match the ports I can see in “ps” for the safe processes. I’ll wait half an hour or so and if I was lucky I guess I should already have some nanos
1 Like
Finally I’ve spotted the error: the docker container redirection didn’t specify protocol and defaulted to TCP. I guess it’ll work when I rebuild the container.
Edit: Working!
3 Likes
Looks good for one node on port 12000
1 Like
Southside said to try it on this Open Port Check Tool - Test Port Forwarding on Your Router
But says not open?
2 Likes
Can’t vouch for that tools as never used it. Have you tried starting a node on port 12000? and watching the logs
Also if that’s your pi you may need to open the fire wall in the pi manually if the fire wall is enabled.
1 Like