Why can't Nvidia support answer all questions in a reply

Why do I have to ask the same question multiple times to get anyone to answer.

They can tell me multiple things in there response, but if my response has multiple questions and results, 9 times out of 10 Nvidia only jumps on the result and ignores other questions in the reply.

It just make information flow very SLOW, and painful to the customer.


Any example for such case? Do you have some topics that no one replies?

If your current problem is not the same question as your original topic, then it is kind of make sense that you need to file a new one. Otherwise, it will make other people painful to read your post.

Also, since everyone here comes from different time zone, if you think your information flow is slow, maybe you can try to share more detail in each comment. Such detail are usually logs or commands but not just some simple comments like “it does not work” or " I cannot change the jetpack version".
If you wonder why other people cannot help you, maybe you can search other posts and see why other users can get help or solutions. Maybe you can refer to how other users file a topic. Also, please keep in mind that if you refuse to change anything or follow any suggestion from those who want to help you, then it has no doubt no issue can get resolved.

[NVIDIA Developer Forums] [Jetson & Embedded Systems/Jetson TX2] R32.3.1.tx2-4g gstream app fails to run nvargus_daemon message

Told the support person I needed information on implementing error handling stuff, and told the support person the results of v4l2-ctl, he jumped on the v4l2-ctl and did not address anything about the messages that he brought up. He can have multiple things in a reply, but ignored my request for more info on the error message handling.

It is also a pain to only get one reply a day, it is almost like the Nvidia people are not in the US… I know some of the Nvidia people are in the US because I have had multiple reply in a day from some support people.

I am going on a year of camera problems and my manager is really upset.


I have had so many topics just stop being replied to, I just give up and start new topics.


Also topics get marked as solution, which are not solutions, it is almost like the support people have to close topics to get paid, you need to pay them based on customer satisfaction. This customer is not satisfied.

I am between a rock and a hard spot, my camera driver is r32.3.1 period, I can’t change because I need to use the camera.

I ask multiple times if I can add debug or collect more info, and it is just crickets.


I think most people know how the product should behave, including the NVIDIA people, but seeing why and how it is not behaving as it should depends on a lot more knowledge, and often actually the ability to reproduce the problem. Most of the NVIDIA people are in the USA, but like most companies, there are not enough people for the number of problems. It wouldn’t matter where they are though if they cannot reproduce the problem and can only guess.

What camera are you working with? If it isn’t one on the list of cameras they have access to, then everything is a guess. However, it doesn’t mean problems for that can’t be solved, but it does mean you will need a lot more information, and there is nothing anyone can do about that. Now imagine that your job was to build a driver for that camera, but you don’t have access to the actual camera…the level of difficulty suddenly goes up a lot. I suppose you could send such a camera to NVIDIA, and perhaps a custom carrier board for testing (though the carrier wouldn’t be important if it comes down to some standard interface, e.g., CSI). Is there a possibility you can send a matching model camera to NVIDIA (I personally lack camera experience and have no test equipment, not even an oscilloscope…mine died, whereas NVIDIA no doubt has some rather advanced test equipment)? That would dramatically aid seeing the problem. NVIDIA would no longer have to “imagine” the camera, they could literally put test equipment on it.

On the other hand, if you want to message me and point out some threads which are current questions, maybe I could see if I could compile an alternate question and format to post (you could private message me a list). You could look it over, and I could keep replying until we come up with a more detailed way of explaining the problems which could be answered. I can’t promise a lot as far as actual camera knowledge since I don’t work on cameras, but I’m betting I could come up with what is needed to actually ask in a way which could be more quickly answered.

For your issue “R32.3.1.tx2-4g gstream app fails to run nvargus_daemon message”. It looks like Jerry already commented in the end. This issue is due to unstable low-level camera driver.
If low-level driver is not stable, it is meaningless to check the upper level application like argus. If you have further questions, you can tag Jerry on that thread and ask what you want to know.
If your camera vendor cannot help you and you don’t know how to get over it, maybe you can ask your vendor to this forum and file topic for their own bug. Sicne they are not our camera partner, we won’t know detail about their sensor unless they come to us.

As for the working time, 95% of support team over jetson forum is not in the US. Even I am not in the US. You just see me in your working time because I check topics in my midnight sometimes. We also assign topics by its category. Thus, maybe those who can handle your topics are all in non-US area so they cannot reply you in your working time. I would say if you want other people to reply your problem immediately, maybe you can reply the topic in their working time too. If you cannot do it, then maybe you should wait for delay response. It is kind of mutual respect. We are waiting for your reply too. It is not only you are waiting.