Re: netsync with port forwarding

Michael Raskin <[email protected]> Sun, 06 Jun 2021 20:21:44 +0200
Newsgroups gmane.comp.version-control.monotone.devel
Message-ID <[email protected]>
>> >Netsync relies on some underlying conventions on the use of TCP for a=20=

>> >two-way connexion.  Is there some other protocol that shares these=20
>> >conventions?  If so I could tell the modem that this other protocol is=
=20
>> >now being used on port 4691.
>>=20
>> I would frankly start with tcpdump on both sides while trying to connec=
t
>> from outside. Routers can break so many things it is not even funny=E2=80=A6
>
>I know.  A port forwarding NAT is an intense kludge.

Static port forwarding doesn't need to be, however routers sometimes=20
have a ton of interesting modes that make things complicated, usually=20
not well named.

Also, it could be that the router port forwards only connections from
outside, while the ISP by default blocks incoming traffic on unknown=20
ports. In the latter case there are two options: actually believing it
is good for safety, and letting through the ports explicitly requested
(if a person can explain what port is needed, this person can probably
be made to clean up their PC if malware gets too annoying for the=20
network); or trying to make residential connections less attractive=20
compared to business connections (doesn't work well in the world of=20
cheap VPS, but=E2=80=A6)

>There was once a publicly accessible site of monotone repositories=20
>called something like mtn-prjk.net -- a kind of github for monotone. =20
>That would have accomplished my desire.  Alas!  it exists no more.

mtn-host.prjek.net, yes=E2=80=A6

>Does netsync support IPv6?
>
>If so there will still be the question of whether the public and the=20
>coffee ships do.)

In principle Monotone even has some code conditional on IPv6 being used.
Among ISPs, both coverage and brokenness vary for IPv6=E2=80=A6