HOWTO: Airspy mini and Airspy R2: Piaware / dump1090-fa configuration

Any idea why the Disc I/O looks like this?

only if dump1090 with native Airspy support is released. without that it runs in net-only mode. So wiedehopf is absolutely right.

Beside that i am pretty sure in these 1682 replies we will find lots of off-topics :grinning:

At least nothing what i see on my fresh installation.
What are you running additionally on the device?

This started after I did play with upgrade the airspy_adsb. On 2 September.
Unfortunately I did also upgrade to Piaware 6.0 about this time, so I am confused.

They said the same about Toyota.

I was about to upgrade to Piaware 6, but I thought I’m making so many adjustments to airspy_adsb it would probably be better to wait than throw another iron into the fire.

Actually i think it is you and foxhunter who are being rude by discussing off topic stuff.
Normally there is no discussion about new development going on in this thread and then i’m happy to discuss disk I/O issues and whatever.
But right now we’re working on improving a very nice program and some people actually like to discuss that as you can see by them posting graphs and so forth.
The discussion being above your head doesn’t necessitate posting irrelevant things.

sudo apt install iotop
sudo iotop -a

That tell you anything about the root cause?
Did you record a sample in memory and not delete it, that could cause full memory, it would be solved by a reboot or deleting the file.

1 Like

Hm, in this case i have no idea. Mine was installed from scratch this morning (dev device) and i started already with Airspy RC2.1 and Piaware 6.0

It's a mystery its a mystery .......
2 Likes

Hah, I see what you did there :smiley:

1 Like

I only gave the reason why i started over again this morning (MLAT issues), nothing to be off-topic. If somebody reads a relation from this, there’s nothing i can do.

back to topic:
the preamble settings are adjusted once in a while, it exceeds the max. CPU setting.
I assume it works as expected?

Sep 04 15:14:17 Raspy airspy_adsb[5466]: CPU 80.0 %, target 90.0 %, adjusting preamble filter: 19 -> 20
Sep 04 15:14:27 Raspy airspy_adsb[5466]: CPU 83.5 %, target 90.0 %, adjusting preamble filter: 20 -> 21
Sep 04 15:14:37 Raspy airspy_adsb[5466]: CPU 85.0 %, target 90.0 %, adjusting preamble filter: 21 -> 22
Sep 04 15:50:48 Raspy airspy_adsb[5466]: CPU 85.0 %, target 90.0 %, adjusting preamble filter: 22 -> 23
Sep 04 16:50:54 Raspy airspy_adsb[5466]: CPU 95.6 %, target 90.0 %, adjusting preamble filter: 23 -> 22

Seems that systemd-journald is the culprit?



It’s not a max but a target, it’ll be happy with ± 5%.

1 Like

Got it thanks.
As the device is feeding only, i leave it with that for the moment

It’s only relative to one core, not the entire device.
As such as long as you’re not losing samples you won’t usually be bottlenecked on that device doing other stuff.

GitHub - wiedehopf/airspy-conf: Configure airspy_adsb for use with readsb or piaware.

You are probably losing lots of samples.
Change your settings so you don’t :slight_smile:

Sep 04 11:47:55 ubuntu systemd[1]: Started Airspy ADS-B receiver.
Sep 04 11:47:55 ubuntu airspy_adsb[2199203]: airspy_adsb v2.2-RC2
Sep 04 11:47:55 ubuntu airspy_adsb[2199203]: Acquired Airspy device with serial 26A464DC28232193
Sep 04 11:47:55 ubuntu airspy_adsb[2199203]: Decoding started at 24 MSPS
Sep 04 11:47:56 ubuntu airspy_adsb[2199203]: Push client connected to localhost:30004 (beast)
Sep 04 11:48:03 ubuntu airspy_adsb[2199203]: CPU 87.9 %, target 95.0 %, adjusting preamble filter: 13 -> 14
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: Dropped samples, adjusting preamble filter: 14 -> 13
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: Dropped samples, adjusting preamble filter: 13 -> 12
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: Dropped samples, adjusting preamble filter: 12 -> 11
Sep 04 12:06:35 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:06:43 ubuntu airspy_adsb[2199203]: CPU 78.8 %, target 95.0 %, adjusting preamble filter: 11 -> 12
Sep 04 12:06:51 ubuntu airspy_adsb[2199203]: CPU 85.7 %, target 95.0 %, adjusting preamble filter: 12 -> 13
Sep 04 12:06:58 ubuntu airspy_adsb[2199203]: CPU 87.8 %, target 95.0 %, adjusting preamble filter: 13 -> 14
Sep 04 12:07:01 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:07:01 ubuntu airspy_adsb[2199203]: Dropped samples, adjusting preamble filter: 14 -> 13
Sep 04 12:07:01 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:07:01 ubuntu airspy_adsb[2199203]: /!\ Lost 98304 samples /!\
Sep 04 12:07:43 ubuntu airspy_adsb[2199203]: CPU 87.6 %, target 95.0 %, adjusting preamble filter: 13 -> 14

I don’t know what happens


No samples lost so far with these seetings. The overall CPU usage is currently below 30% constantly on a Pi4 (not overclocked)

none on mine 
Sep 04 15:22:16 rpi4-20210823 systemd[1]: Started Airspy ADS-B receiver.
Sep 04 15:22:16 rpi4-20210823 airspy_adsb[15750]: airspy_adsb v2.2-RC2
Sep 04 15:22:16 rpi4-20210823 airspy_adsb[15750]: Listening for beast clients on port 47787
Sep 04 15:22:16 rpi4-20210823 airspy_adsb[15750]: Acquired Airspy device with serial 62CC68FF218F5F17
Sep 04 15:22:16 rpi4-20210823 airspy_adsb[15750]: Decoding started at 12 MSPS
Sep 04 15:22:17 rpi4-20210823 airspy_adsb[15750]: Push client connected to localhost:30004 (beast)
Sep 04 15:22:26 rpi4-20210823 airspy_adsb[15750]: CPU 47.0 %, target 75.0 %, adjusting preamble filter: 4 -> 6
Sep 04 15:22:36 rpi4-20210823 airspy_adsb[15750]: CPU 61.1 %, target 75.0 %, adjusting preamble filter: 6 -> 7
Sep 04 15:22:46 rpi4-20210823 airspy_adsb[15750]: CPU 68.2 %, target 75.0 %, adjusting preamble filter: 7 -> 8
Sep 04 15:22:56 rpi4-20210823 airspy_adsb[15750]: CPU 62.3 %, target 75.0 %, adjusting preamble filter: 8 -> 10
Sep 04 15:23:06 rpi4-20210823 airspy_adsb[15750]: CPU 69.6 %, target 75.0 %, adjusting preamble filter: 10 -> 11
Sep 04 15:23:55 rpi4-20210823 airspy_adsb[15750]: CPU 67.8 %, target 75.0 %, adjusting preamble filter: 11 -> 12
Sep 04 15:24:55 rpi4-20210823 airspy_adsb[15750]: CPU 69.1 %, target 75.0 %, adjusting preamble filter: 12 -> 13
Sep 04 15:25:55 rpi4-20210823 airspy_adsb[15750]: CPU 69.3 %, target 75.0 %, adjusting preamble filter: 13 -> 14

Use a target of 85 that will work or set -e directly without -C.
-C 90 might work 
 just check the log.