version 2.25.00 suddenly takes up 100% resources

[DG2MDJ]

I have the same Problem.
My PC:
Ryzen 5 5600G with Radeon Graphics, 16GB Ram
Windows 11 22H2
Log4OM V 2.25.0.0
WSJT-X V2.5.4

After about 6 hours runtime the CUP usage quickly ramps up to 100%
I had this about 3 times with this PC always after 6 to 7 hours runtime.

[N5ZY]

This weekend I created a new W10 virtual machine in VirtualBox on Ubuntu. I installed MySQL 8, Visual C++ 2015-2022 Redistributable, Trusted QSL and Log4OM 2.25.0.0.
This VM is used for logging only and JTDX is sending it UDP packets on port 2237.
(I may have 3 radios sending log entries to it if i'm on HF, 6m, and VHF).
WSJTX and JTDX are running on a dedicated physical computer also running Ubuntu.

Overnight with nothing being logged at all Log4OM on the brand new VM began to consume 100% CPU. It seems to be creating lots of threads for clr.dll!DLLRegisterServerInternal+0x4ad60. Normally while idle it uses 18 threads but when I found the CPU in the VM at 100% this morning it had 218 threads.

VirtualBox_2023 W10_22h2_08_01_2023_04_31_29-CPU100pct_cropt.png (193.97 KiB) Viewed 1078 times

[IW3HMH]

Thanks for details, i will investigate this behaviour.
About voacap: every hour log4om recalculates the whole voacap data set (all bands on every 4digit gridsquare in the world, pretty intensive activity).
those data are used to extimate propagation info on cluster realtime view and draw the propagation maps (it doesn't affect the realtime calculation made on every callsign typed)

This could effectively drain a lot of resources.
Thread proliferation is, indeed, not normal and worth a check

Thanks

[IW3HMH]

This weekend I created a new W10 virtual machine in VirtualBox on Ubuntu. I installed MySQL 8, Visual C++ 2015-2022 Redistributable, Trusted QSL and Log4OM 2.25.0.0.

Can you reply the same situation but without any application sending packets to Log4OM? (no JT packets)
This will help me isolate the potential issue

tnx
Daniele IW3HMH

[SQ4BJA]

Hi,
In my case it helps. Without JTDX LOG4OM works with below 1% CPU usage. I also change Hamlib to Omnirig. It also reduced the CPU usage. Now back to LOG4OM testing with JTDX - I'll check if hamlib was the problem.

Tom SQ4BJA

EDIT: Unfortunately. After a few hours of use, the CPU usage went up to 100% again. Eliminating Hamlib helped but didn't solve the problem.

[N5ZY]

Can you reply the same situation but without any application sending packets to Log4OM? (no JT packets)
This will help me isolate the potential issue

Will do. I have both VM's running (I downgraded my new w10vm to 2.24.0.0 so I could log contacts this weekend and 2.24 worked well).
The old VM still has 2.25.0.0 so it is now started without packets from FTDX (the FTDX machine is reporting to the IP of the new w10vm).
I will report findings tomorrow.

[N5ZY]

Can you reply the same situation but without any application sending packets to Log4OM? (no JT packets)
This will help me isolate the potential issue

After 7 hours without data from WSJT/FTDX, the thread count looks normal and CPU is normal.
2.25.0.0 has 21 threads, CPU idle
2.24.0.0 has 18 threads, CPU idle.

After 19 hrs, v2.25 it without any activity on UDP ports it has 21 threads, CPU idle.
After 24 hrs. v2.24 has 17 threads, CPU idle, also without UDP traffic.