I have a Ci20 connected via an unmanaged 10/100 switch) which is in turn connected to a wifi router, where my laptop is connected via wifi.
If I run "nuttcp -S -4" on my laptop (nuttcp is a network benchmarking tool) and then run "nuttcp -4 -i1 [my IP]" I'll have the following output;
ci20@ci20:~$ nuttcp -4 -i1 192.168.8.177
9.3750 MB / 1.00 sec = 78.6434 Mbps 0 retrans
9.8125 MB / 1.00 sec = 82.3118 Mbps 0 retrans
9.9375 MB / 1.00 sec = 83.3630 Mbps 0 retrans
nuttcp-t: v6.1.2: Error: server not ACKing data
ci20@ci20:~$ Timeout, server ci20 not responding.
So the Ci20 stops seeing my nuttcp server, but can still see me, and then it falls over and I can't ping or route to the Ci20, even after rebooting the wifi router and power cycling the switch.
I don't think the Ci20 has frozen up - I see to much activity on the LEDs, and when it does freeze up (heavy CPU load) those lights are full on but have no activity.
This is with Debain 8 (non-beta).
Curiously, this is the first time I'm testing a new switch. Prior to this I had a little Cisco desktop switch - which I began to suspect, and over which I had large performance issues with a PINE64. I switched it to a TP-Link, and now I connect perfectly to the PINE64 - but now the Ci20 has the problem described above when using nuttcp (and it did not before).