I have been testing a pair of zedx’s w/ a dockerized zed sdk for the past little while now and have noticed that our largest issue so far has been intermittent argus failures and we are not clear on what exactly is causing this problem, but it’s becoming a pretty big bottleneck for us to reliably build applications that use the zedx cameras.
we normally have minimal problems when we first start using the cameras, but after about an hour of continuous use we begin to experience the argus issues/failures.
for reference, we built our docker images using the examples and information from the zed_ros2_wrapper (so we have the relevant volume mounts and permissions) and also use a capture card from stereolabs with 2 ZEDX cameras. we primarily use the latest zed-ros-wrapper (for ROS1). we use a 32gb orin agx running jp5.0.2
below is a screenshot of the error log that continuously shows up whenever we try testing after the initial failure occurs.
long continuous runs. For example, I would start up all of my nodes and the cameras would be running fine for about an hour or so and then fail and pretty much all subsequent attempts at starting up again after that initial crash are met with the argus failure logs
we originally started using 4.0.2 along with the driver version that was initially released (I can confirm the exact version later today). We recently upgraded the SDK to 4.0.5, but are still using the same driver
I tried updating to v0.5.4 driver today, but it appears that I am still facing the previous issues. I have attached the nvargus logs in case they might be insightful (we see a lot of errors regarding wrong frequency and it crashed around when the invalid state errors started popping up). Is there anything else we can try?
Hi @woudie-swap
can you write an email to support@stereolabs.com to open a ticket and get direct tech support?
We will investigate your problem in detail.
