On 11/10/23 22:30, Gregory Nowak via Dng wrote:
> On Fri, Nov 10, 2023 at 09:11:57PM -0600, Michael S. Keller via Dng wrote:
>> From the log, it looks like TP-Link may use some variant of
>> OpenWRT, as there is a sign that it contains the Luci administration tool.
>
> Not surprised. Most if not all TP-Link routers seem to use an older
> version of OpenWRT. That may prove useful in diagnosing this from the
> gateway end if you're able to get SSH access to the OpenWRT console.
>
>>
>> As for using Wireshark, I can see when a big transfer stalls, but am not
>> sure what useful bits to glean from the results. When a big Samba file
>> transfer stalls, I see IPv6 router advertisements and a bunch of other
>> stuff. Eventually there's a TCP reset, and soon after the file transfer
>> resumes.
>
> Can you install ndisc6 on this host if you don't have it installed
> already? What do you get if you run:
>
> rdisc6 ethx
>
> when the connection times out, where ethx is your network interface?
>
> Greg
>
>

Hop limit                 :           64 (      0x40)
Stateful address conf.    :          Yes
Stateful other conf.      :          Yes
Mobile home agent         :           No
Router preference         :       medium
Neighbor discovery proxy  :           No
Router lifetime           :         1800 (0x00000708) seconds
Reachable time            :  unspecified (0x00000000)
Retransmit time           :  unspecified (0x00000000)
  Source link-layer address: 48:22:54:50:90:10
  from fe80::4a22:54ff:fe50:9010

This appears to be the same as rdisc6 output at other times.

This test was similar to my earliest tests - start netcat listening, and
feed it a spew from /dev/zero.

It stall about 9 seconds into the tcpdump capture, and doesn't recover
until 125 seconds in.

There are some TCP Spurious Retransmissions and Dup ACKs in that span,
along with a lot of router solicitation and router advertisements