files.netgate.com IPv6 down?

[Squuiid]

I have IPv6 configured and I'm unable to connect to files.netgate.com for updates.

/root: ping6 www.netgate.com
PING(56=40+8+8 bytes) 2a01:4b00:8351:xxxxxxxxxxxxxxx --> 2606:2c40::c73c:671e
16 bytes from 2606:2c40::c73c:671e, icmp_seq=0 hlim=249 time=3.512 ms
16 bytes from 2606:2c40::c73c:671e, icmp_seq=1 hlim=249 time=3.494 ms
16 bytes from 2606:2c40::c73c:671e, icmp_seq=2 hlim=249 time=3.495 ms

/root: ping6 files.netgate.com
PING(56=40+8+8 bytes) 2a01:4b00:8351:xxxxxxxxxxxxxxx --> 2610:160:11:18::209
^C
--- files.netgate.com ping statistics ---
16 packets transmitted, 0 packets received, 100.0% packet loss

Anyone else seeing this?

[johnpoz]

@Squuiid not seeing a problem now.

[24.11-RELEASE][admin@sg4860.home.arpa]/root: ping 2610:160:11:18::209
PING(56=40+8+8 bytes) 2001:470:snipped::2 --> 2610:160:11:18::209
16 bytes from 2610:160:11:18::209, icmp_seq=0 hlim=52 time=47.690 ms
16 bytes from 2610:160:11:18::209, icmp_seq=1 hlim=52 time=58.902 ms
16 bytes from 2610:160:11:18::209, icmp_seq=2 hlim=52 time=54.002 ms
16 bytes from 2610:160:11:18::209, icmp_seq=3 hlim=52 time=46.528 ms
16 bytes from 2610:160:11:18::209, icmp_seq=4 hlim=52 time=55.070 ms
16 bytes from 2610:160:11:18::209, icmp_seq=5 hlim=52 time=46.965 ms

Or from here.

[Squuiid]

@johnpoz Thanks for checking. From there is indeed correct.
I seemingly have an ISP routing issue with Hyperoptic in the UK. Damn, their support is awful. Not looking forward to trying to get them to address this.
Meanwhile pfSense is unable to check for updates as it prefers IPv6 over 4 by the looks of it.

[patient0]

@Squuiid said in files.netgate.com IPv6 down?:

Meanwhile pfSense is unable to check for updates as it prefers IPv6 over 4 by the looks of it.

You can change that under System / Advanced / Networking

-> Prefer IPv4 over IPv6

Not sure if a reboot is necessary or not.

[Squuiid]

@patient0 Temporarily I guess. Kinda don’t want to do it globally but yeah, that should work. Thanks.

It would be nice if the upgrade routine would fail back to IPv4 instead but I guess this is quite a rare circumstance. Wouldn’t hurt though.

[patient0]

@Squuiid said in files.netgate.com IPv6 down?:

I guess. Kinda don’t want to do it globally but yeah, that should work.

Mmmh, never tried it but what happens if you create an IPv4 host override (or add it to /etc/hosts) for whatever DNS name(s) is(are) used (only files.netgate.com?)?

[Squuiid]

@patient0 Very good suggestion, thank you, but unfortunately it didn't work. I guess it still prefers IPv6, checks that it resolves, and it does, so then insists on using it despite not being routable.
(Use local DNS, ignore remote checked)

Edit: Given how seemingly poor my ISP's IPv6 implementation is it's probably best I do just prefer IPv4 globally and be done with it. This fixes the issue immediately and has the side benefit of improved latency.

[patient0]

@Squuiid I don't know how much time you can or want to invest into this but I fear there are more DNS entries involved, check out Unable to upgrate 23.09.1 to 24.11.

If you run the update from the comand line and add the entries directly to /etc/hosts or maybe ~/.netrc you can monitor right away what happens.

Edit: wow, the RTT and RTTsd for the IPv4 are amazing, are you living in their basement? :) and I thought mine aren't too bad

[Squuiid]

@patient0 Haha, yeah, but look how bad my IPv6 is!
Going to leave it with IPv4 preferred, but thank you very much for taking the time to help. Those were great suggestions, and in the end this issue just highlighted how bad my ISP's IPv6 is.
Hyperoptic, get your IPv6 house in order!