"Alexander V. Makartsev" <avbe...@gmail.com> wrote: > On 22.10.2024 08:17, Max Nikulin wrote: > > On 22/10/2024 03:21, Alexander V. Makartsev wrote: > >> If I manually throttle these connections they disconnect after > >> some time and soon after a new connection from another IP from the > >> same subnet or different network establishes. > > > > May it happen that their internet providers have NAT and pools of > > IP addresses for outgoing connections? New connection just uses > > another IP from a pool with no efforts at the client side. > > > > A shot in the dark: maybe client settings are rather aggressive in > > respect to peer number to connect, but their bandwidth are > > saturated by other (local) peers. > > > Connection IPs span over different networks, so I don't think it is a > pool or a few subnets of a single ISP. > Here is a few example IPs I gathered from those suspicious > connections: 36.32.56.219 > 36.32.63.210 > 36.106.178.254 > 36.106.54.166 > 112.101.176.215 > 121.56.211.154 > 182.245.68.120 > 222.211.26.158 > 117.181.164.206 > 182.136.100.183 > 59.34.152.170 > 144.0.15.230 > 163.142.241.158
whois on those addresses gives several real-looking personal email addresses. I'd be inclined to send a polite email to some/all of them asking if they know about these connections and can explain them. > I've already accumulated pretty long list. They all point to > different ISP networks in China. > The only thing I'm certain of is that they use "bttracker.debian.org" > to get peer information. > Maybe this is somehow tied to "webseed peer" of > "debian-12.5.0-amd64-netinst.iso" torrent? > I don't know enough about torrent trackers or webseeds to be able to > tell. > > Like I said before, I also seldom get normal torrent connections from > China IPs, and they behave like the rest peers from around the world. > They report correct information and status about themselves, request > chunks they need to download, up to 100% of completion and then > disconnect.
