This reminded me about attic and crawlspace wiring
I changed the network cable on my server and connected it directly to my router. I did not remember that I had it going into a switch and then to the router. They are both next to each other and the switch was needed as I ran out of ports on the router.
The results have been great. I've run the test for a few days and just stopped it. I had three network errors. I will change out the other cables at the turners to see if that resolves.
I do see transport errors and a few sequence errors but I will chalk those up to the extremely windy weather we have been having. I am fine with some signal issues occasionally but not the playback stops I was seeing before.
Good to hear. Troubleshooting sporadic issues like that can be a real PITA.
Start by watching side by side on the HDHomeRun app and Channels, preferably with just one HDHomeRun on to eliminate a variable.
Thanks for this great tool for troubleshooting.
You may want to update your post on the SD Forum to let them know that this is what you used to test.
https://info.hdhomerun.com/info/troubleshooting:network_packet_loss#troubleshooting_network_packet_loss
I'm also curious what the sequence errors it reports are.
I assume it's one of these and my guess is missed packet counter (jump in sequence numbers)
https://info.hdhomerun.com/info/hdhomerun_config#checking_the_signal_strength
seq = symbol error quality (number of uncorrectable digital errors detected)
miss = missed packet counter (jump in sequence numbers)
That's it. The packet Continuity Counter
There is a counter on each packet that runs from 0-15, and within each data stream the packets climb sequentially from 0 to 15 then start over at 0. Any time that the numbers don't go up in perfect order is a sequence error.
-- Video statistics --
144277043 packets received, 0 overflow errors, 4 network errors, 56 transport errors, 1 sequence errors
I can't remember when I started this running but I just stopped it and these are the results. Sample size is ~144 million packets. I'm sure ideally it should be all zeros but is this terrible?
I'm not the expert, you should ask Silicondust.
4 network errors in ~144 million packets doesn't seem bad.
The transport and sequence error numbers are also pretty low (although those have to do with the received broadcast signal and not your network).