This has me thinking that this is a PHY-layer issue.
How long are your Ethernet runs, and are you using Cat5e cabling?
On 2016-05-03 12:59, John Shields wrote:
It's possible that we're dealing with a memory leak somewhere. Could you watch the memory consumption of simple_ra over time?
Also, could you just run something like uhd_fft with the same sample rate for long periods to see if it gets the same 'D' treatment?
On 2016-05-03 04:32, John Shields wrote:
Ok, Marcus, will do. Overnight (NZ time) I got 1 D with the Atheros. I should have said that the N200 was in the shack and the Ubuntu system in the house. I have several runs of Ethernet out to the shack so will do some experimentation on the different connections. Then I will move the Ubuntu system to the shack and see if colocation helps. These investigations will take some days but will let you know.
Kind Regards,
John
|