Solved Netcup vserver not reachable

[mro]

the symptoms look like https://forums.freebsd.org/threads/netcup-ipv6-intermittant.96489/ but the cure doesn't.

From outside:

$ ping 2a03:4000:5:f44:affe:beef::1
PING(56=40+8+8 bytes) 2001:4090:a242:835b:5ee0:c5ff:fe7f:6fed --> 2a03:4000:5:f44:affe:beef:0:1
^C
--- 2a03:4000:5:f44:affe:beef::1 ping statistics ---
848 packets transmitted, 0 packets received, 100.0% packet loss


Outgoing ping and ssh are fine.

From inside the ipv6-only FreeBSD 14.3-p9 vServer:

$ cat /etc/rc.conf
cloudinit_enable="YES"
qemu_guest_agent_enable="YES"
hostname=v2202602340009436687.bestsrv.de
# ipv6_activate_all_interfaces="YES"
# ipv6_defaultrouter=fe80::1%vtnet0
# ipv6_ifconfig_vtnet0=2a03:4000:5:f44:7442:baff:fe2d:d8e1
ipv6_defaultrouter="2a03:4000:5::2"
# my ip range is 2a03:4000:5:f44::/64
ifconfig_vtnet0_ipv6="inet6 2a03:4000:5:f44:affe:beef::1/48"


I understood to limit the range from 64 to 48 but am not confident with the ipv6_defaultrouter. How can I learn more?

 

[SirDice]

I understood to limit the range from 64 to 48

This extends the range, not limit it. Smaller prefix -> bigger range. Similar to IPv4, a /16 is a bigger subnet (i.e. more host addresses) than a /24.

 

[mro]

uh. My office machine doesn't but another vServer at a different hoster does get a ping.

So I suppose the problem is on my office side. My office network also takes ages to reach the 'other' host by ssh. Hm.

 

[zapata]

Why /48? By default you get a /64? Check the netcup ccp.
Why "2a03:4000:5::2"? Try: ipv6_defaultrouter="fe80::1%vtnet0"

 

[mro]

Why /48? By default you get a /64? Check the netcup ccp.
Why "2a03:4000:5::2"? Try: ipv6_defaultrouter="fe80::1%vtnet0"

my whole journey started with https://paulwilde.uk/ponderings/my-battles-with-netcup-ipv6-on-freebsd/ - there is a recommendation of a 48 range and the defaultrouter address.

I returned to the defaults and things look ok.