I have successfully set up Piaware 3.x and it seems to be sending data to FA on ADS-B however after three days I’m not sending any MLAT data. I went onto the site today and ensure my receiver position and height was set up and it seems as though it is. Do I have to wait to see MLAT data appearing or is something else wrong here?
Also, I cannot find the ‘send data to receiver’ button the update process talks about. Does that only appear when new updates are ready to send?
Thanks for that. I’ve tried moving the position pin a little and it seems to have picked up some MLAT data now, we’ll see if that fixes the anomaly.
The upgrade info is on the ADS-B pages. http://flightaware.com/adsb/piaware/upgrade select Piaware 3 and the info tells you to use the ‘send command to device’. I don’t see that option on my ABS-B page.
It is important to get an accurate position or mlat won’t work properly; make sure you zoom right in when selecting a position (or enter coordinates from a GPS)
The upgrade info is on the ADS-B pages. http://flightaware.com/adsb/piaware/upgrade select Piaware 3 and the info tells you to use the ‘send command to device’. I don’t see that option on my ABS-B page.
Ah, “send command to device” is correct, I was wondering where you were seeing “send data to receiver”…
Are you using a sdcard install or a package install? There is a bug with package installs in 3.1.0 that can make the send command option disappear.
I’ve run piaware 2.x and dump1090-mutability both compiled on a beaglebone black running wheezy for over 1 year. I recently switched to jessie and to the piaware 3 release and dump1090-fa. The moment I did this, I went from reporting between 50000 and 70000 MLAT positions per day to around 200 to 300. I am getting the timing anomaly message on my page. What changed here? I only upgraded because I was getting emails and banners on the web page asking me to. Has Piaware 3 added so much stress to the Beaglebone that it can’t handle it anymore? I’d really like to go back to the old setup.
OK, so that’s the probable cause. I don’t know what the kworker thread is doing, but I’d guess you took a kernel upgrade going from wheezy to jessie which triggered something there. Perhaps USB related if you are also having mlat timing problems; dropping USB data will cripple mlat.
dump1090-fa generally will take a bit more CPU than the older dump1090, and it will put a bit more load (+20%) on the USB subsystem as it runs at a higher sampling rate, but I would not expect that to make much of a difference unless the system was right on the edge previously. You could compile an older version of dump1090 if you wanted to try that; but the tradeoff is that the demodulator in older dump1090s is not as good, so you will receive less traffic, and older dump1090s are not mlat-aware so you will not be able to use them to display mlat results.
Thank you Oliver, I agree. I have reverted to 2.1-5 and 1.2-3 on wheezy for now. CPU load is about 21% fro fa-mlat-client and 28% for dump1090. Total rarely breeches 50% and that is with the local feed page open. kworker processes are wayyy down the list. I did try to install the 3.1 package on a wheezy image, but several dependencies were a bit too old.
I am trying to work at it from the kworker process end. There is loads of info about high usage kworker processes hogging cpu on all different architectures.
Just an addendum, being a glutton for punishment I upgraded to 3.3 through the menu pulldown menu on the “my adsb” page. Everything fine, CPU under 50%, so it seems likely that this is something related to Jessie. I won’t mark solved since I think I technically hijacked the thread and my situation probably doesn’t apply to the OP. Thanks for the feedback obj! And thanks for all the work you do!