I have a wifi router and an extender in my home and I believe that there might be some misconfiguration in the extender. I often have high latency or lower than expected throughput when connecting to the extender.
My main question is:
How do I troubleshoot a wifi extender that seems to cause a lot of duplications?
ANSWER: In my case, the problem was a faulty NIC on the PC ("d"). See EDIT3 for a full explanation and look at the comments for valiable tips on WIFI thoubleshooting.
What I have tried so far is using ping, mtr, traceroute/tracepath. I'll next describe my scenario and then give some example ping outputs:
Scenario
What I have is:
a. wire-only, ISP-provided router connected to the Internet
b. TP-Link Archer-C60 connected to "a" (the ISP router) via ethernet
c. TP-Link AC 750 (RE200) connected to "b" via a 5.8Ghz wifi channel
d. Home PC connected to "c" via 2.4Ghz wifi channel
No other device uses the 5.8Ghz channel. Other devices use either b's or c's 2.4Ghz wifi channel to access the network.
The physical model of the environment is:
Legend:
-- and | are walls
' is a ethernet cable
a is the ISP-provided router
b is the TP-Link Archer-C60 wifi router
c is the TP-Link AC 750 (RE200) wifi extender
d is a home PC
|--------------|
| d |
|c |
-----------| |------ ------| |
| |
| b|------- --- -| |
| '| |--------------|
| '|
| a|
| |
|-------------|
Example pings
I have ran ping from host "d" while no other device is connected to the network. When doing this, I can easily get 5 to 7 duplicates for a same ping reply. Some example outputs:
Pinging b (192.168.0.1) from d:
--- 192.168.0.1 ping statistics ---
421 packets transmitted, 421 received, +487 duplicates, 0% packet loss, time 420022ms
rtt min/avg/max/mdev = 1.053/10.543/188.536/21.423 ms
Pinging 8.8.8.8 from d:
--- 8.8.8.8 ping statistics ---
421 packets transmitted, 420 received, +867 duplicates, 0% packet loss, time 419810ms
rtt min/avg/max/mdev = 35.480/47.917/256.235/21.228 ms
EDIT1: When running tcpdump, I get lots of ICMP time exceeded in-transit. Small sample, where 192.168.0.1 is "b" and 192.168.0.100 is "d"...
EDIT2:
removed
ICMP time exceeded in-transitlog, added in EDIT1, as it was unrelated.virtual topology of the environment
The virtual topology is:
a --- cat5 --- b >>> 5.8Ghz >>> c ))) 2.4Ghz ))) d and other hosts devices
U
U
2.4Ghz
U
U
other hosts devices
That is:
- Internet is provided by 'a' to 'b' via cat5 and distributed by 'b' via wireless
- router 'b' and extender 'c' communicate via the 5.8Ghz channel
- hosts connect to either 'b' or 'c' via the 2.4Ghz channel (theoretically they could connect via the 5.8Ghz channel, but I do not have compatible devices at the moment)
EDIT3:
It turns out, it was a faulty PCI wireless NIC on my PC ("d"). The NIC was dropping a lot of packets and sending lots of duplicates. This behavior clogged the repeater and eventually the whole network, affecting other devices. I replaced the NIC and now everything works just fine.
To identify this, I simply got a pair of notebooks and attempted to test connectivity inside and outside the network. Also, I did all tests with "d" turned on and then repeated them with it turned off. The difference was very noticeable.
Connectivity inside the network: I used iputils' ping to test dropping rate while flooding (-f flag). Drop rate was near zero between the two notebooks, but was around 20% to my PC. The PC also exhibited a lot of DUPs. tcpdump showed duplicated packets as belonging from my PC or only going towards my PC. When my PC was turned on, ping between the two devices had a few DUPs and drops. When the PC was turned off, no DUPs or drops were identified.
Connectivity outside the network: I used mytraceroute, fast.com, and speedtest.net to check connectivity outside the network. The latency was very high and throughput very low when my PC was turned on. The network behaved perfectly, with very low latency and near nominal throughput, when my PC was turned off.
With the new NIC, everything is working like a charm.
