Since yesterday I am having a new "fun" problem in my LAN:
Pinging the salt fiber box too often from the LAN leads to packet loss. This includes DNS, ICMP and all routed traffic.
Pinging from my box fails after two packages:
PING 10.0.0.1 (10.0.0.1) 56(84) bytes of data. 64 Bytes von 10.0.0.1: icmp_seq=1 ttl=64 Zeit=467 ms 64 Bytes von 10.0.0.1: icmp_seq=2 ttl=64 Zeit=2.91 ms (--- no more responses ---)
Now pinging other machines beyond the LAN are no longer reachable.
Meanwhile, other machines can ping both, the gateway and others beyond. Their traffic is not affected in any shape or form.
The salt box appears to have a list of hosts that the box does not agree with. Since I had to disable the fiber boxes WiFi already after my attempt of fixing the salt fiber box have failed all my traffic reaches the fiber box over Ethernet.
The affected boxes can no longer reach the admin interface of the salt fiber box.
The worst part is: not all traffic is blocked, but most. That means that browsing the web is broken for the most part (eg. big pictures and videos) but mail still somewhat works.
Some applications like signals can not cope with a faulty link and loose cohesion when ~90% of traffic is dropped.
Since the effects appear limited to the MAC & IP address combination, I temporarely solved it by changing my MAC address and receiving a new dynamic IP.
Others like SSH or protocols tunneling over it like git or socks5 get very slow and sometimes fail to establish a handshake.
I hope salt will fix this issue in the future with a firmware fix:
Model Name: Salt_Fiber_Box Firmware Version: v1.01.38 build105 Boot Code Version: 0.00.01 Hardware Version: R01B