TCP connections to Azure Front Door IPv6 endpoints in the 2620:1ec::/32 range fail completely from Orange Poland's IPv6 network (AS5617). ICMP probes successfully reach Azure's internal infrastructure, confirming the issue is not a BGP peering problem — packets enter Azure's network but TCP sessions to the Front Door anycast IPs never complete. Multiple production sites using Azure Front Door are affected.
| DNS name | CNAME chain | Azure Front Door IPv6 |
|---|
TCP connections to Azure Front Door IPv6 endpoints in the 2620:1ec::/32 range fail completely from Orange Poland's IPv6 network (AS5617). ICMP probes successfully reach Azure's internal infrastructure, confirming the issue is not a BGP peering problem — packets enter Azure's network but TCP sessions to the Front Door anycast IPs never complete. Multiple production sites using Azure Front Door are affected.
I've managed to compile a working firmware with this config: https://github.com/gleber/qmk_firmware/commit/54078ec73c0ff2bd01984242f748a837c1954d03
This isn't tested much, just the basics: keys and split setup. LED, bootloader magic key and any other features are not yet implemented. Tested only the Miryoku layout so far.
I've disassembled it and found a few things.
I was trying to configure native ipv6 on my NixOS box, which is running in online.net's datacenters. They provide you
a /48 or smaller subnet and a DUID which is used during DHCP. In dhcpcd vocabulary DUID is called clientid.
online.net's help page about IPv6 configuration was not very useful, since NixOS uses dhcpcd by default. The page which allowed me to make it all work was https://community.online.net/t/tutorial-ipv6-with-dhcpcd/3804
It looks like the only really required entries in networking.dhcpcd.extraConfig are interface and static though.
I assumed that clientid is also necessary, but it looks like dhcpcd cares much more about content of /etc/dhcpcd.duid.
Of course replace enp2s0 with the name of the main interface.
I hereby claim:
To claim this, I am signing this object:
| import Language.Haskell.TH | |
| f n m = do | |
| nm1 <- newName "x" | |
| let b = replicate (m - 1) WildP | |
| a = replicate (n - m) WildP | |
| return (LamE [TupP (b++[VarP nm1]++a)] (VarE nm1)) |
| 1> F1 = future:new(fun() -> timer:sleep(3000), 10 end). | |
| {future,<0.36.0>,#Ref<0.0.0.1660>,undefined} | |
| 2> F2 = future:new(fun() -> timer:sleep(3000), 5 end). | |
| {future,<0.39.0>,#Ref<0.0.0.2014>,undefined} | |
| 3> F2:get() * F1:get(). | |
| 50 |
| # -*- mode: Makefile; fill-column: 80; comment-column: 75; -*- | |
| ERL = $(shell which erl) | |
| ERLFLAGS= -pa $(CURDIR)/.eunit -pa $(CURDIR)/ebin -pa $(CURDIR)/*/ebin | |
| REBAR=$(shell which rebar) | |
| ifeq ($(REBAR),) | |
| $(error "Rebar not available on this system") |
| %% Analysis results: | |
| { analysis_options, | |
| [{callers, false}, | |
| {sort, own}, | |
| {totals, true}, | |
| {details, false}]}. | |
| % CNT ACC OWN | |
| [{ totals, 6985519,124011.127,1236847.232}]. %%% |
| gleber@first:~/code/opensource$ git clone git@github.com:gleber/exat.git exat | |
| Cloning into exat... | |
| remote: Counting objects: 240, done. | |
| remote: Compressing objects: 100% (148/148), done. | |
| remote: Total 240 (delta 125), reused 202 (delta 87) | |
| Receiving objects: 100% (240/240), 705.27 KiB | 172 KiB/s, done. | |
| Resolving deltas: 100% (125/125), done. | |
| gleber@first:~/code/opensource$ cd exat | |
| gleber@first:~/code/opensource/exat$ make deps all | |
| ./rebar get-deps |