Correct. In short, once again: 1. The below works fine for a normal relay with internal class-A IPv4 and a reachable public IPv4 / IPv6 - it also works fine for the "Relay part" of a bridge. ORPort 7.8.9.10:1234 NoListen ORPort 10.0.0.7:1234 NoAdvertise ORPort [2345:bad:face::]:1234 2. Using the line "ServerTransportListenAddr obfs4 10.0.0.7:4321" (i.e. binding the internal class-A IPv4) for an obfs4 bridge has the following effects: - The respective bridge line for obfs4 using the external IPv4 works well, you can immediately use the bridge line in TBB - In metrics, the bridge at first appears with transport "none" and https://bridges.torproject.org/status?id=<hashed fingerprint> indicates a working "vanilla" bridge w/o plugable transport - A while later (one day or so), metrics updates to plugable transport "obfs4" in "green", but https://bridges.torproject.org/status?id=<hashed fingerprint> show a dysfunctional obfs4 bridge! - The bridge distribution method shown in metrics stays forever as "none", even specifying a different bridge distribution in torrc has no impact 3. Using the line "ServerTransportListenAddr obfs4 [2345:bad:face::]:4321" (i.e. specifying the externally reachable IPv6) has the following effects: - The respective bridge line for obfs4 using the IPv6 works well, you can immediately use the bridge line in TBB - In metrics and https://bridges.torproject.org/status?id=<hashed fingerprint>, the bridge is shown as "functional" - The bridge distribution method shown in metrics stays forever as "none", even specifying a different bridge distribution in torrc has no impact HOWEVER - the "observation time window" may be "too short" / "impacted by an independent issue": We've had in the recent past a couple of times the issue, that the distribution method for bridges wasn't updated for more than one week ! See e.g. this bridge E80088962F97FE190D734D67E0C656C2A6EADEE7, which is an ordinary webtunnel bridge without any special config, (re-)created by me mid of last week. It still shows "none" as distribution method and therefore does not get any traffic at all. Also in other cases, manually setting the bridge distribution method in torrc on other bridges does not seem to have any impact right now. So it could very well be, that there is a general issue with bridge distribution method not being updated, which in fact has nothing to do with the above observations? @meskio - is there a general issue with bridge distribution not being updated again? So to summarize again, we have two topics: A. Not possible to properly advertise an obfs4 bridge bound to an internal class-A IPv4 B. Potential/unclear issue of bridge distribution not being updated, which either is an unrelated general issue currently affecting all bridges, or specifically related to the above described observations. Regards, C. Am 06.08.2025 um 23:59, "Gary C. New via tor-relays" <tor-relays@lists.torproject.org> schrieb: C: To confirm... Configuration of the ORPort with Internal NoAdvertise IPv4 and External IPv6 Advertise enables the Bridge Status functional? However, the issue with the Bridge Distribution Mechanism remaining "None" is still not a viable solution? Respectfully, G On Tuesday, August 5, 2025 at 12:30:03 PM MDT, C. Pe. via tor-relays <tor-relays@lists.torproject.org> wrote: Hello all, I can give an update: This example config for an OBFS4 Bridge behind IPv4-NAT using secondary IPv4+IPv6 seems to fully work (as you may have noticed, the example values of external IPv6 and IPv6 obviously are nonsense): BridgeRelay 1 ORPort 7.8.9.10:1234 NoListen ORPort 10.0.0.7:1234 NoAdvertise ORPort [2345:bad:face::]:1234 ServerTransportPlugin obfs4 exec /usr/bin/obfs4proxy ServerTransportListenAddr obfs4 [2345:bad:face::]:4321 ExtORPort auto ContactInfo johndoe@nowhere.org Nickname MyCoolBridge As you can see, ... - it uses Non-NAT secondary IPv6 for the bridge line - The Address statement previously discussed isn't necessary. HOWEVER: When I use the below statement instead (tested with a brand-new bridge) - hence the class-A internal IP for the port binding: ServerTransportListenAddr obfs4 10.0.0.7:4321 the following can be observed: 1. Using the bridge line "obfs4 7.8.9.10:4321 <fingerprint> cert=<cert> iat-mode=0" in the TOR Browser works perfectly fine 2. The bridge in metrics.torproject.org is shown with Transport protocols "none" ! 3. The page https://bridges.torproject.org/status?id=<hashed fingerprint> shows "vanilla IPv4: functional" The real-life example of my test is hashed fingerprint 50ACF3A9AF5A7980F4CFE6CF5F9B27E56B2E72EB So my question to the experts: - Is an IPv4 via NAT using a secondary IPv4 simply not foreseen/considered with obfs4proxy? - Or is there a special (unknown to me) setting, maybe via "ServerTransportOptions obfs4 <...?...>" in torrc to be set? Looking forward to your reply. (Depending on it, I may have to convert it to an obfs4 IPv6 bridge) Thanks & regards, C. Am 31.07.2025 um 22:09, "C. Pe. via tor-relays" <tor-relays@lists.torproject.org> schrieb: Many thanks for the reply. Thanks for the hint with the status check - it shows what I was suspecting already, that my bridge is "dysfunctional", despite the config outlined in my previous mail.
I think you can reduce this by just: ORPort 1234 Not exactly: "ORPort 1234" results in the below bindings: 0.0.0.0:1234 [::]:1234
whilst ORPort 10.0.0.7:1234 NoAdvertise ORPort [2345:bad:face::]:1234 result in the below, intended bindings, which leaves those posts free for the other IPs: 10.0.0.7:1234 [2345:bad:face::]:1234 And when I omit the "ORPort 7.8.9.10:1234 NoListen" statement, then - despite the "Address 7.8.9.10" statement - I get the log line below: External address seen and suggested by a directory authority: <the default IPv4 address, which I don't want to use> Those statements work perfectly well for a relay. But to me it seems, that the obfs4 transport has a problem: * obfs4 IPv4: dysfunctional Error: Bridge address is not valid Last tested: 2025-07-30 16:23:09.776330892 +0000 UTC m=+115200.406981514 (26h41m55.597470102s ago) It seems that NAT in combination with a secondary IPv4 creates issues with the obfs4 transport. As said, my (intended) bridge line works in my browser, but BridgeDB can't parse the proper bridge (i.e. obfs4) address, maybe it parses the class-A private IP (which obviously won't work at all)? I assume that an IPv6 obfs4 should work (I'll try now), but maybe there is a ServerTransportOptions setting dealing with the needed IPv4 mapping? Or is the combo NAT plus secondary IPv4 simply not possible for obfs4proxy ? Regards, C. Am 31.07.2025 um 13:12, meskio via tor-relays <tor-relays@lists.torproject.org> schrieb:Quoting C. Pe. via tor-relays (2025-07-30 22:17:02)
Dear experts,
I have a special question, for which I could not find 100% clear documentation after a research.
Situation: - VPS with more than one IPv4 and IPv6 address - IPv4 within the VPS is private class-A only (i.e. 10.x address), so port binding happens to the private Class-A
- External IPv4 is linked outside the VPS to the private Class-A - of course, VPS is externally reachable by the external IPv4
- I would like to use a secondary IPv4, i.e. NOT the one, which is shown, when you run the command 'curl -4 icanhazip.com' from within the VPS) - I would like to use a secondary IPv6, i.e. NOT the one, which is shown, when you run the command 'curl -6 icanhazip.com' from within the VPS)
Let's assume, for my example only, I have the following secondary IP addresses : - external IPv4 7.8.9.10 - internal IPv4 10.0.0.7 - IPv6 [2345:bad:face::]
Let's further assume, my ORPort was 1234 and the obfs4 port was 4321
In order to have TOR distribute the correct bridge line "obfs4 7.8.9.10:4321 <...>", is the following torrc config correct?
BridgeRelay 1
ORPort 7.8.9.10:1234 NoListen ORPort 10.0.0.7:1234 NoAdvertise ORPort [2345:bad:face::]:1234
I think you can reduce this by just: ORPort 1234 As I believe Adress and OutboundBindAddress will do the rest for you. But others here might have a better torrc-fu than me.
Address 7.8.9.10 Address [2345:bad:face::] OutboundBindAddress 10.0.0.7 OutboundBindAddress [2345:bad:face::]
ServerTransportPlugin obfs4 exec /usr/bin/obfs4proxy ServerTransportListenAddr obfs4 10.0.0.7:4321
ExtORPort auto ContactInfo johndoe@nowhere.org Nickname MyCoolBridge
With this config, I indeed am able to connect to the bridge using the "obfs4
7.8.9.10:4321 <...>" bridge line,
but is this line also communicated like that by the bridge distributor, or do I need to adapt the above config to ensure proper distribution?
I think this configuration is correct. The published address will be the one set in the Address. You can always check if the bridgeline distributed by rdsys is correct by visiting: https://bridges.torproject.org/status?id=<your hashed or not fingerprint> This should display one bullet point per IP/transport, so in your case two, one for IPv6 and another for IPv4. Take into account that it might take up to 4h for this page to be updated when you do changes. -- meskio | https://meskio.net/ -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- My contact info: https://meskio.net/crypto.txt -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Nos vamos a Croatan. _______________________________________________ tor-relays mailing list -- tor-relays@lists.torproject.org To unsubscribe send an email to tor-relays-leave@lists.torproject.org -- Sent with https://mailfence.com Secure and private email _______________________________________________ tor-relays mailing list -- tor-relays@lists.torproject.org To unsubscribe send an email to tor-relays-leave@lists.torproject.org -- Sent with https://mailfence.com Secure and private email _______________________________________________ tor-relays mailing list -- tor-relays@lists.torproject.org To unsubscribe send an email to tor-relays-leave@lists.torproject.org _______________________________________________ tor-relays mailing list -- tor-relays@lists.torproject.org To unsubscribe send an email to tor-relays-leave@lists.torproject.org -- Sent with https://mailfence.com Secure and private email