PF client dead after upgrade

@nsmits Are you able to get a backtrace with GDB? Should be a simple case of typing "bt" and hitting return.

Sorry if I am teaching you to suck eggs there!
Code:
Thread 15 "pfclient" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb11fd460 (LWP 7698)]
dictionary_get (d=0x0, key=key@entry=0xef09c "flightpaths", def=def@entry=0x0)
    at /var/pfclient/PFClient-NG/pfclient/pfclient/vendor/iniparser/dictionary.c:185
185     /var/pfclient/PFClient-NG/pfclient/pfclient/vendor/iniparser/dictionary.c: Bestand of map bestaat niet.
(gdb) bt
#0  dictionary_get (d=0x0, key=key@entry=0xef09c "flightpaths", def=def@entry=0x0)
    at /var/pfclient/PFClient-NG/pfclient/pfclient/vendor/iniparser/dictionary.c:185
#1  0x00043ee8 in pf_web_server_handle_request (request=request@entry=0xb2c005d8)
    at /var/pfclient/PFClient-NG/pfclient/pfclient/pf_web_server.c:507
#2  0x00044004 in pf_web_server_request_parse_completion_callback (parser=<optimized out>)
    at /var/pfclient/PFClient-NG/pfclient/pfclient/pf_web_server.c:931
#3  0x000461f4 in http_parser_execute (parser=parser@entry=0xb2c005f8, settings=settings@entry=0xb11fce18,
    data=data@entry=0xb11e8e18 "GET /ajax/aircraft HTTP/1.1\r\nUser-Agent: Wget/1.20.1 (linux-gnueabihf)\r\nAccept: */*\r\nAccept-Encoding: identity\r\nHost: 127.0.0.1:30053\r\nConnection: Keep-Alive\r\n\r\n", len=len@entry=161)
    at /var/pfclient/PFClient-NG/pfclient/pfclient/vendor/httpparser/http_parser.c:1871
#4  0x00044108 in pf_handle_web_request (socket_desc=0xb38007d8) at /var/pfclient/PFClient-NG/pfclient/pfclient/pf_web_server.c:237
#5  0xb6f81494 in start_thread (arg=0xb11fd460) at pthread_create.c:486
#6  0xb6e6b578 in ?? () at ../sysdeps/unix/sysv/linux/arm/clone.S:73 from /lib/arm-linux-gnueabihf/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

please let me know if you need more info or help
 
Thank you very much, I can replicate this now. I think we'll be able to get a fix tested today and a new release out next week.

Out of interest... it looks like you are using something to poll the `/ajax/aircraft` endpoint of the client.
  • Do you know what is triggering this?
  • Also if you manage to stop it does the client not crash any more?
 
Can anyone that was having issues please try:

Code:
http://client.planefinder.net/pfclient_5.0.152_armhf.deb

Normal install method, and please let me know how you get on with that.

Thank you!
 
Code:
pi@flightradar:~ $ systemctl status pfclient.service
● pfclient.service - LSB: planefinder.net ads-b decoder
   Loaded: loaded (/etc/init.d/pfclient; generated)
   Active: active (running) since Fri 2021-02-19 15:14:10 CET; 27min ago
     Docs: man:systemd-sysv-generator(8)
    Tasks: 7 (limit: 3860)
   CGroup: /system.slice/pfclient.service
           └─28113 /usr/bin/pfclient -d -i /var/run/pfclient.pid -z /etc/pfclient-config.json -y /var/log/pfclient $

feb 19 15:14:10 flightradar systemd[1]: Starting LSB: planefinder.net ads-b decoder...
feb 19 15:14:10 flightradar pfclient[28107]: Starting pfclient: pfclient.
feb 19 15:14:10 flightradar systemd[1]: Started LSB: planefinder.net ads-b decoder.

also working for me :D
 
Hi Lee

My PF Radar Receiver did an auto update to client 5.0.152 about 15.18 but since then it keeps closing down and restarting

Any suggestions?
 
Hi Nigel,

The update process for Plane Finder Radar units is a little different so sometimes they can reboot a few times. Seems to be ok now though?
 
Successfully installed and running OK on following:

pfclient_5.0.152_armhf.deb

  1. RPi Model 2B / Piaware SD Card image 4.0 (Buster)
  2. OrangePiPC / Armbian Buster
  3. RPi Model 4 / Raspberry Pi OS 64-bit (Buster)

pfclient_5.0.152_i386.deb
  1. Kali Linux 2020 amd64 on PC (AMD Ryzen 3, 3.60 GHz CPU)
  2. Ubuntu 20.04.1 x86_64 on PC (AMD Ryzen 3, 3.60 GHz CPU)

pfclient_5.0.152_i386.tar.gz
  1. Arch Linux amd64 on PC (AMD Ryzen 3, 3.60 GHz CPU)
 
Last edited:
By the way, new interface:
Now I can see the ugly white dots on the trail of the planes. My eyes almost hurt, so I have to make a better antenna ... :)
Which antenna you are currently using? I use Flightaware 1090 MHz antenna, and still get white dots ;)

1613906201110.png
 
Last edited:
Right now, I use an FA stick, yesterday I climbed up to try a Franklin. Both of them have pros and cons.

Of course, there will always be blind spots and spots with a weaker reception of signals. So far they have been, though so far they have not been so obvious.

Just recently, I simulated countless variants of a "stacked-phased" antenna. Very delicate construction, but maybe I managed to model a promising piece:
stacked farfield.jpg

The lowest lobe starts from 1° thus it MUST be vertical. Really.
 
Last edited:
Here is an even better one - but it behaves hysterically even because of a height difference of up to 30-40 centimeters.
stacked.jpg

I hope that white spots would appear a bit farther than now.. This is my next antenna project. 9 dBd is expected, counting with the effect of mast, cables and other things.
 
Last edited:
15.66 dBi !!!!
What is the design of antenna?
Don't bother with the big number. The reference gain is excessive due to the old HAM program.
The actual difference is not that great. As I mentioned about 9 dBd (gain over the dipole) is expected. As for the modern calculations, it is about 11-12 dBi. Depending on the position of mast, - backwards 6 dBd and either forward or sides can have 9 dBd gain. Really depends on little things.

This antenna contains 4 half-wave (normal) dipoles - fed one by one.
They have to be tuned in their place because they are really sensitive at the ends. (this is why we can force them to change the caracteristics) The optimal distance between the current-maximum points and the effect of the mast must be experimented with in order to achieve the correct result. What you need to pay close attention to is that they have similar jx values, close to zero.
Outer and inner pairs are "brothers", so they are pairs in 2 step combination...

I used the old MMANA-GAL prog

We need two 2-way power combiners --> combined with each other again. (it is third) The simplest way is building the set into a single box. The whole thing is coax pieces and resistors between the connectors... thus a diy version is preferred. Piece of cake. It is a wilkinson combiner, but I have to use them in two steps because of the different pairs of dipoles. (2 of them have 1 neighbour only)

If you want, you can play with NEC (I do not have it)
numbers.jpg

The last wire is the mast.
R= radius
Material: Cu wire



Note: In this construction, Reactance is more important than Impedance. A phase deviation above 10-15 degrees is already unacceptable if you are really striving for quality. Use real ground in modeling, play carefully with height.

3D plot
3D.jpg


Mast has to be in the model, since it will carry the coax cables vertically and also the Wilkinson combiner in metal box. It has to reach the center of the upmost dipol. Combiner shall be at the center-line of the antenna, on the opposite side on the mast. After combining the segments, must there be a choke at the botom-line of the antenna - on the mast. (2-thread coil on the coaxial cable towards the amplifier.)
Dipoles must be in phase, this is why we use the same lenght of coax for each.
structure.jpg

I wonder what results you will get with another program.
 
Last edited:
Don't bother with the big number. The reference gain is excessive due to the old HAM program.
The actual difference is not that great. As I mentioned about 9 dBd (gain over the dipole) is expected. As for the modern calculations, it is about 11-12 dBi. Depending on the position of mast, - backwards 6 dBd and either forward or sides can have 9 dBd gain. Really depends on little things.

This antenna contains 4 half-wave (normal) dipoles - fed one by one.
They have to be tuned in their place because they are really sensitive at the ends. (this is why we can force them to change the caracteristics) The optimal distance between the current-maximum points and the effect of the mast must be experimented with in order to achieve the correct result. What you need to pay close attention to is that they have similar jx values, close to zero.
Outer and inner pairs are "brothers", so they are pairs in 2 step combination...

I used the old MMANA-GAL prog

We need two 2-way power combiners --> combined with each other again. (it is third) The simplest way is building the set into a single box. The whole thing is coax pieces and resistors between the connectors... thus a diy version is preferred. Piece of cake. It is a wilkinson combiner, but I have to use them in two steps because of the different pairs of dipoles. (2 of them have 1 neighbour only)

If you want, you can play with NEC (I do not have it)
View attachment 3104
The last wire is the mast.
R= radius
Material: Cu wire



Note: In this construction, Reactance is more important than Impedance. A phase deviation above 10-15 degrees is already unacceptable if you are really striving for quality. Use real ground in modeling, play carefully with height.

3D plot
View attachment 3105

Mast has to be in the model, since it will carry the coax cables vertically and also the Wilkinson combiner in metal box. It has to reach the center of the upmost dipol. Combiner shall be at the center-line of the antenna, on the opposite side on the mast. After combining the segments, must there be a choke at the botom-line of the antenna - on the mast. (2-thread coil on the coaxial cable towards the amplifier.)
Dipoles must be in phase, this is why we use the same lenght of coax for each.
View attachment 3106
I wonder what results you will get with another program.

Waiting for you to build it, put it on field test, and post results.
 
Ok, I will build one. Maybe in next month.
I am not allowed to put antennas to (outside) the roof.
Only attic - under the tiles - is available. We'll see if it works for a comparison....
 
Last edited:
Back
Top