[–] 7 points 1 week ago (2 children)

Pretty sure that's a load-reduction measure due to AI scraping. Git history produces a ton of links which scrapers traverse through and generate load. I'm pretty sure just cloning the repo with git over https works fine.

  • source
  • parent
  • context
  • [–] 2 points 4 months ago* (last edited 4 months ago) (1 child)

    DHCP usually provides three things:

    • IP Address (+ subnet mask)
    • Defaulft gateway IP
    • DNS servers

    In the DHCP lease you for example get the IP 192.168.10.42/24. That IP is assigned to your interface which creates a implizit route like 192.168.10.0/24 dev netinterface. Now you know how to reach other hosts in your local network.

    In the same DHCP lease, you should also receive a default gateway - likely 192.168.10.1. this would add a route like default via 192.168.10.1.

    Without the default route, packets for IPs other than your local network never leave your host because it doesn't know where to send them to.

    DNS servers are also provided, though before troubleshooting that, I would just ping addresses like 8.8.8.8 (google dns) for testing connectivity.

    (Btw, what I'm trying to say is: DHCP is working, but your server configuration may not be correct for what you're trying to do)

    You could also skip troubleshooting DHCP by running ip route add default via 192.168.10.1 to test if your firewall is configured correctly. That should enable you to ping other networks.

  • source
  • parent
  • context
  • [–] 1 point 4 months ago (3 children)

    If that's the only route then your issue starts there. If you want to reach anything other than the local network there needs to be a default route to the gateway.

    The default route is sent along the dhcp lease so my suspicion is that the issue is in the dhcp server config.

  • source
  • parent
  • context
  • [–] 10 points 4 months ago

    Just for context, this technology is not made for any specific end-consumer. This is for backbone connections of ISP and higher tier networks.

    In that space, 400Gb/s is the current stuff that's being rolled out with 800Gb/s starting to get traction. 1.2Tb is just the next step to reduce the amount of Fibers and Ports used per connection (renting fibers of an undersea cable is expensive).

    Though one issue I see slowing adoption is that new fibers are needed. With coherent optics we already achieve 800G with the same single mode fibers already in use.

  • source
  • [–] 5 points 1 year ago (1 child)

    That's not necessarily very easy. These certs would have to show up in public certificate transparancy logs for most browsers to accept them. If this happens on a government scale it would surely get noticed, though the question remains what you're left to do if the government forces it anyways...

    See https://en.m.wikipedia.org/wiki/Certificate_Transparency section "Mandatory certificate transparency"

  • source
  • parent
  • context
  • [–] 3 points 1 year ago*

    If by RJ45 you mean erhernet over standard 8P8C (i.e. RJ45) your looking for SFP-T modules. Can't connect that to your DSL line though since that's not Ethernet.

    There do seem to exist G.fast sfp modules for dsl, but do note that sadly sfp module and router brand incompatibility is sadly a thing.

    Related note: At our company we also had quite some issues getting some of these more special sfp modules working (e.g. XGS-PON), so ymmv with a dsl one.

    (If you're looking for a device to plug sfp modules into, I can second MikroTik here. They're also great about accepting 3rd party SFPs)

  • source
  • parent
  • context
  • view more: next ›