Arch Linux

gdkags

Member

Registered: 2010-10-12

Posts: 18

systemd-networkd routing broken?

Hi!

I've got six containers running through machinectl (default configurations), which I can reach with no problems via link-local ipv6.
Nginx seems to prefer ipv4 tho, so that's when the problem surfaced:

machinectl:

MACHINE   CLASS     SERVICE        OS VERSION ADDRESSES 
doku      container systemd-nspawn -  -       10.0.0.5… 
hash      container systemd-nspawn -  -       10.0.0.66…
miniflux  container systemd-nspawn -  -       10.0.0.53…
nextcloud container systemd-nspawn -  -       10.0.0.86…
searx     container systemd-nspawn -  -       10.0.0.90…
solr      container systemd-nspawn -  -       10.0.0.44…

ip route:

10.0.0.0/28 dev ve-searx proto kernel scope link src 10.0.0.1 
10.0.0.16/28 dev ve-miniflux proto kernel scope link src 10.0.0.17 
10.0.0.32/28 dev ve-solr proto kernel scope link src 10.0.0.33 
10.0.0.48/28 dev ve-nextcloud proto kernel scope link src 10.0.0.49 
10.0.0.64/28 dev ve-doku proto kernel scope link src 10.0.0.65 
10.0.0.80/28 dev ve-hash proto kernel scope link src 10.0.0.81

I mean, 'doku' is living in 10.0.0.5 but the route is set to 10.0.0.64/28 so this can not work.

Is systemd-networkd's routing broken or am I missing something crucial? Just wanted to check, before I submit a bug report.