RE650 v2 - unstable!! (fixed, see my reply)
RE650 connected to Deco main over wireless (2.4ghz)
Updated to latest firmware.
Ping time is mostly 5ms. Mostly. App diagnostics say it is in an ideal location!
Broadcasts a local network on both 2.4 and 5 using lowest power.
Testing throughput:
2.4 is about 30 mbit to iperf locally
5 ghz is about 80 mbit to iperf locally
but in 5ghz mode it regularly (every 5 or 10 minutes) just blanks out.
in 2.4ghz it is more reliable but still blanks out. Pings literally just timeout. After a while, I get a reconnection...
It doesn't get too hot. The room is not too hot.
The blackouts happen and any device (Linux, IOS or OSX) lose connection.
The local signal strength is -37 as reported by OSX where 0 is max and -100 is worst.
Here is an example, it happned while I was typing this.
64 bytes from 10.10.10.1: icmp_seq=1042 ttl=64 time=7.763 ms
64 bytes from 10.10.10.1: icmp_seq=1043 ttl=64 time=5.787 ms
64 bytes from 10.10.10.1: icmp_seq=1044 ttl=64 time=8.473 ms
64 bytes from 10.10.10.1: icmp_seq=1045 ttl=64 time=11.406 ms
64 bytes from 10.10.10.1: icmp_seq=1046 ttl=64 time=7.607 ms
64 bytes from 10.10.10.1: icmp_seq=1047 ttl=64 time=6.589 ms
64 bytes from 10.10.10.1: icmp_seq=1048 ttl=64 time=10.840 ms
64 bytes from 10.10.10.1: icmp_seq=1049 ttl=64 time=8.551 ms
64 bytes from 10.10.10.1: icmp_seq=1050 ttl=64 time=5.706 ms
64 bytes from 10.10.10.1: icmp_seq=1051 ttl=64 time=5.233 ms
64 bytes from 10.10.10.1: icmp_seq=1052 ttl=64 time=9.618 ms
64 bytes from 10.10.10.1: icmp_seq=1053 ttl=64 time=6.540 ms
Request timeout for icmp_seq 1054
64 bytes from 10.10.10.1: icmp_seq=1055 ttl=64 time=6.207 ms
64 bytes from 10.10.10.1: icmp_seq=1056 ttl=64 time=114.521 ms
Request timeout for icmp_seq 1057
64 bytes from 10.10.10.1: icmp_seq=1058 ttl=64 time=6.653 ms
Request timeout for icmp_seq 1059
64 bytes from 10.10.10.1: icmp_seq=1060 ttl=64 time=5.631 ms
64 bytes from 10.10.10.1: icmp_seq=1061 ttl=64 time=506.421 ms
64 bytes from 10.10.10.1: icmp_seq=1062 ttl=64 time=8.827 ms
Request timeout for icmp_seq 1063
64 bytes from 10.10.10.1: icmp_seq=1064 ttl=64 time=5.677 ms
64 bytes from 10.10.10.1: icmp_seq=1065 ttl=64 time=12.118 ms
Request timeout for icmp_seq 1066
Request timeout for icmp_seq 1067
Request timeout for icmp_seq 1068
ping: sendto: No route to host
Request timeout for icmp_seq 1069
ping: sendto: No route to host
Request timeout for icmp_seq 1070
ping: sendto: No route to host
Request timeout for icmp_seq 1071
Request timeout for icmp_seq 1072
64 bytes from 10.10.10.1: icmp_seq=1073 ttl=64 time=439.407 ms
64 bytes from 10.10.10.1: icmp_seq=1074 ttl=64 time=16.729 ms
64 bytes from 10.10.10.1: icmp_seq=1075 ttl=64 time=20.013 ms
Note that it isn't just latency spiking from noise. "NO ROUTE TO HOST".
This will trigger OSX to try to connect to the next available known network.
The whole wifi connection goes dead, and reconnection then happens.