[Lightning-dev] Dropping Tor v2 onion services from node_announcement
laolu32 at gmail.com
Wed Sep 22 23:03:50 UTC 2021
Earlier this week I was helping a user debug a Tor related issue, and
realized (from the logs) that some newer Tor clients are already refusing to
connect out to v2 onion services.
On the lnd side, I think we'll move to disallow users creating a v2 onion
service in our next major release (0.14), and also possibly "upgrade" them
to a v3 onion service if their node supports it. I've made a tracking issue
I ran a naive script to gauge how much of the network is using Tor
generally, and also v2 onion services extract the following stats:
num nodes: 12844
num tor: 8793
num num v2: 66
num num v3: 8777
This counts advertised addresses total, so it likely over estimates, so you
can treat this as an upper bound. Thankfully only 60 or so v2 addresses seem
to be even _advertised_ on the network, so I don't think this'll cause much
Another interesting tidbit here is that: _over half_ of all advertised
addresses on the network today are onion services. I wonder how the rise of
onion service usage (many nodes being tor-only) has affected: e2e payment
latency, general connection stability, and gossip announcement propagation.
In terms of actions we need to take at the spec level, it's likely enough to
amend the section on addrs in the node_announcement message to advise
implementations to _ignore_ the v2 addr type.
On Tue, Jun 1, 2021 at 3:19 PM darosior via Lightning-dev <
lightning-dev at lists.linuxfoundation.org> wrote:
> Hi all,
> It's been almost 9 months since Tor v2 hidden services have been
> The Tor project will drop v2 support in about a month in the latest
> release. It will then be entirely be dropped from all supported releases by
> More at https://blog.torproject.org/v2-deprecation-timeline .
> Bitcoin Core defaults to v3 since 0.21.0 (
> https://bitcoincore.org/en/releases/0.21.0/) and is planning to drop the
> v2 support for 0.22 (https://github.com/bitcoin/bitcoin/pull/22050),
> which means that v2 onions will gradually stop being gossiped on the
> Bitcoin network.
> I think we should do the same for the Lightning network, and the timeline
> is rather tight. Also, the configuration is user-facing (as opposed to
> Bitcoin Core, which generates ephemeral services) which i expect to make
> the transition trickier.
> C-lightning is deprecating the configuration of Tor v2 services starting
> next release, according to our deprecation policy we should be able to
> entirely drop its support 3 releases after this one, which should be not so
> far from the October deadline.
> Opinions? What is the state of other implementations with regard to Tor v2
> Lightning-dev mailing list
> Lightning-dev at lists.linuxfoundation.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Lightning-dev