Using Redhawk SDR 2.10 in a Centos7 Docker Container, I have created a node that contains among other things a GPP device configured with default property values. When I launch the domain and select my node as the device manager, I start seeing logs that indicate the GPP is busy because the max open files threshold is exceeded. You will notice in the logs below that the "threshold" is larger than the "actual":
2023-10-03 14:03:51 INFO BlackGPP:2608 - send_threshold_event: OPEN_FILES threshold exceeded (resource_id=ulimit threshold_value=1017118.75 measured_value=19)
2023-10-03 14:03:51 INFO BlackGPP:1922 - GPP BUSY, REASON: ULIMIT (MAX_FILES) Threshold: 1048576 Actual: 19
2023-10-03 14:03:54 INFO BlackGPP:1938 - GPP BUSY, (first/duration: 2023-Oct-03 14:03:51.485157/00:00:03.034449) REASON: ULIMIT (MAX_FILES) EXCEEDED Threshold: 1048576 Actual: 19
2023-10-03 14:03:57 INFO BlackGPP:1938 - GPP BUSY, (first/duration: 2023-Oct-03 14:03:51.485157/00:00:06.065073) REASON: ULIMIT (MAX_FILES) EXCEEDED Threshold: 1048576 Actual: 18
2023-10-03 14:04:00 INFO BlackGPP:1938 - GPP BUSY, (first/duration: 2023-Oct-03 14:03:51.485157/00:00:09.094847) REASON: ULIMIT (MAX_FILES) EXCEEDED Threshold: 1048576 Actual: 18
I do not remember this occurring when using Redhawk 2.6 in the same way. I tried setting property thresholds::files_available=0; however,this had no effect. Playing with gpp_limits properties causes a PartialReconfigurationException. If I set the property thresholds::ignore=true, the problem goes away.