One of my Servers regularly crashes with Error:
malloc(): unsorted double linked list corrupted
Config:
[server]
port=8073 # Server port
html_root="frontend/dist/" # HTML files to be hosted
otherusers=1 # Send where other users are listening, 0 to disable
threads=2
[websdr]
register_online=true # If the SDR should be registered on https://sdr-list.xyz then put it to true
name="Sardinia SDR - Perfugas (SS)" # Name that is shown on https://sdr-list.xyz
antenna="Wellbrook ALA1530" # Antenna that is shown on https://sdr-list.xyz
grid_locator="JN40LT" # 4 or 6 length Grid Locatlr shown on https://sdr-list.xyz and for the Distance of FT8 Signals
hostname="sardinia-sdr.ddns.net" # If you use ddns or something to host with a domain enter it here for https://sdr-list.xyz
[limits]
audio=100
waterfall=100
events=100
[input]
sps=60000000 # Input Sample Rate
#fft_size=262144 # small samplerates
#fft_size=1048576 # medium
#fft_size=2097152 # high
fft_size=8388608 # very high
fft_threads=2 # Default = 4
brightness_offset=0 # Waterfall brightness offset. Reduce to negative if you see black regions in the waterfall
frequency=0 # Baseband frequency
signal="real" # real or iq
accelerator="opencl" # Accelerator: none, cuda, opencl
audio_sps=12000 # Audio Sample Rate 12000 is needed for FT8 decoding
audio_compression="flac" # flac or opus
waterfall_size=2048 # Default=2048
waterfall_compression="zstd" # zstd or av1
smeter_offset=5
[input.driver]
name="stdin" # Driver name
format="s16" # Sample format: u8, s8, u16, s16, u32, s32, f32, f64
[input.defaults]
frequency=14317000 # Default frequency to show user
modulation="USB" # Default modulation
Startup
sudo rx888_stream/target/release/rx888_stream -f ./rx888_stream/SDDC_FX3.img -s 60000000 -g 90 -m low -r --pga -o - | sudo build/spectrumserver --config config_rx888.toml
Is your websdr running out of samples to process?
Or it the CPU too weak to run that many threads?
What CPU are you using?
What does 'dmesg' tell you when it happens?
Has the RX888 been modified thermally? If not the USB-chip gets too hot and start giving errors, but worse in a few days it will die.
All the rest looks fine to me.
CPU (Intel(R) Core(TM) i5-6600T CPU @ 2.70GHz), RAM and Thermal of RX888 should be fine. Maybe something in here, some usb powersave or so.
[ 10.138207] rfkill: input handler disabled
[99408.358208] perf: interrupt took too long (2514 > 2500), lowering kernel.perf _event_max_sample_rate to 79500
[235580.890666] loop19: detected capacity change from 0 to 560352
[530443.064864] Lockdown: systemd-logind: hibernation is restricted; see man ker nel_lockdown.7
[530571.324422] usb 1-13: USB disconnect, device number 7
[530571.642137] usb 2-8: new SuperSpeed USB device number 2 using xhci_hcd
[530571.663201] usb 2-8: New USB device found, idVendor=04b4, idProduct=00f1, bc dDevice= 0.00
[530571.663214] usb 2-8: New USB device strings: Mfr=1, Product=2, SerialNumber= 3
[530571.663222] usb 2-8: Product: RX888mk2
[530571.663227] usb 2-8: Manufacturer: sdr prototypes
[530571.663233] usb 2-8: SerialNumber: 0009002301961B06
[532222.517209] perf: interrupt took too long (3148 > 3142), lowering kernel.per f_event_max_sample_rate to 63500
[537719.827435] perf: interrupt took too long (3940 > 3935), lowering kernel.per f_event_max_sample_rate to 50750
[546755.782514] perf: interrupt took too long (4932 > 4925), lowering kernel.per f_event_max_sample_rate to 40500
[568118.764324] perf: interrupt took too long (6275 > 6165), lowering kernel.per f_event_max_sample_rate to 31750
[682590.087282] usb 2-8: USB disconnect, device number 2
[682590.299249] usb 1-13: new high-speed USB device number 8 using xhci_hcd
[682590.447932] usb 1-13: New USB device found, idVendor=04b4, idProduct=00f3, b cdDevice= 1.00
[682590.447938] usb 1-13: New USB device strings: Mfr=1, Product=2, SerialNumber =3
[682590.447941] usb 1-13: Product: WestBridge
[682590.447943] usb 1-13: Manufacturer: Cypress
[682590.447944] usb 1-13: SerialNumber: 0000000004BE
[682590.552699] usb 1-13: USB disconnect, device number 8
[682590.919307] usb 2-8: new SuperSpeed USB device number 3 using xhci_hcd
[682590.940232] usb 2-8: New USB device found, idVendor=04b4, idProduct=00f1, bc dDevice= 0.00
[682590.940238] usb 2-8: New USB device strings: Mfr=1, Product=2, SerialNumber= 3
[682590.940240] usb 2-8: Product: RX888mk2
[682590.940242] usb 2-8: Manufacturer: sdr prototypes
[682590.940244] usb 2-8: SerialNumber: 0009002301961B06
[682596.528559] Lockdown: systemd-logind: hibernation is restricted; see man ker
Seems your Kernel doesn't like it...perf errors, where it tells you it hoggs the system....
Try adding this to /etc/sysctl.conf
kernel.perf_cpu_time_max_percent=0
Just add it at the end and reboot.
It's a protection mechanism....see if it helps.
Get also this error sometimes. Seems not related to usb as I use only gigabit ethernet input.
Will run in gdb soon to track the bug.
Ill also look at it soon again but gdb didnt really help me.
Quote from: F5OEO on Nov 23, 2024, 12:29 PMGet also this error sometimes. Seems not related to usb as I use only gigabit ethernet input.
Will run in gdb soon to track the bug.
Did you apply the Chat-box crash fix? See top post in this section of the forum.
Maybe I am fortunate, but I have not had any crashes that were the fault of the PhantomSDR+ application. The one crash I have had was when I was moving the PhantomSDR+ apps to use systemctl and did not kill all processes that were already running. That was my fault. And I also have not made the chat-box mod because I was not having crashes. So with that I thought I'd post how I built my system and maybe there is something in there that is different.
My CPU = Xeon E5-1620 v3 @ 3.50GHz
My GPU = NVIDIA Corporation GP106 [GeForce GTX 1060 3GB]
System RAM = 16G
Storage is a 250G SSD
I begin with installing Ubuntu 22.04.5 Server.
I apply all the updates and go no further than Ubuntu Server 22.
I install the NVIDIA drivers for my GPU and remove any Ubuntu drivers for that card that was installed during the
initial install.
I verify OpenCL is working for my GPU.
After these steps are accomplished, then I run the auto install on Steven's github and correct any errors that I see on the compile. Check that screen closely and look over any dependencies that may fail and fix those before moving on and something else I do is rm -rf the PhantomSDR-Plus directory path before starting another compile.
I fix the dependencies, then try a fresh run.
So far, so good. I have had no crashes and I am seeing sometimes near 80 users on the weekends.
Running gdb for hours, and no crash at all. It will not help tracking the bug !