Brain OS image for staging environment

support
1

Hi farm_ng team,

We’re progressing well on our implementation of a LiDAR scanning app to run on the Amiga Brain, but we’re finding some issues when we go from our staging environment (a VM) to the Brain itself. There are a host of things that might be getting in the way, but in the interest of trying to make our staging environment as close to the Brain itself as possible, would it be possible for us to get a copy of the Linux image that contains the current version of the Brain OS?

Cheers,
George

2

Hey George,

Sorry the late response. Unfortunately we are not able to public share the image for the Brain. That would not only be an intelectual property problem, but also would make it easier for people to bypass safety features we add in our OS.

Anything in particular you are having trouble with?

3

Thanks Marcelo, and no problem, I understand.

We’re having some issues with the LiDAR driver seg-faulting when we move the services to the Brain itself. I’m not working on it directly anymore, we’ve partnered with a local engineeering firm, but some of the details are here: Segmentation fault (core dumped) Error · Issue #376 · SICKAG/sick_scan_xd · GitHub

We’re just trying to rule out all of the possible causes (hardware, OS etc.). It looks like we might have had some progress based on the last post in that GitHub Issue, though.

4

I read through the thread and looks like you have a solution! Thanks for sharing it here!

#LIDAR #SICK

5

Hi @MScatolin,

Thanks for following up, I’m working with George and the team on this project.

As mentioned, thankfully the segmentation fault issue with the sick drivers has been resolved with the un-optimised build.

We’re still experiencing an issue with the sick driver tcp thread exiting unexpectedly. See the thread here for more details.

It occurs when we’re execute an unrelated section of code (data processing, numpy array operations etc). Unfortunately were hitting a bit of a road-block in determining who/what is exiting the sick tcp thread. We suspect maybe the OS? Are you aware of any watchdogs or OS processes that could cause this?

The system resources (Memory/CPU) seem ok. It does run without issues on a different platform (different HW & OS).

Appreciate any help or guidance in troubleshooting this.

Thanks!

6

Hey guys,

I would start checking journalctl on the services we run (you can find the name on Settings → Services tab of the brain. Note the user we run them is adminfarmng, so something like sudo -u adminfarmng journalctl -fa --user-unit farmng-launcher-backend.service.

Now, that said, my colleagues can have more insight at this, so I will ask them to check here as well.

7

Hey guys,

I did ask about this with our SW team and we are not aware of anything in the base OS or any of our services that would directly interfere with the TCP connection.

One idea was to stop our services and reduce the traffic on the switch to amke sure its not us? We have a script for that at /farm_ng_image/brain2/stop_services.sh.