Strange APT repository behavior

I have setup three Xavier NX boxes using the SDK Manager. All three are running the latest jetpack 5.0.2. All three are identical in their network and apt configuration. For some reason, doing a sudo apt update works fine on two of the boxes, but the third one keeps saying that it cannot find a route to repo.download.nvidia.com:443. The specific message looks like the following:

Hit:1 http://ports.ubuntu.com/ubuntu-ports focal InRelease
Get:2 http://ports.ubuntu.com/ubuntu-ports focal-updates InRelease [114 kB]
Get:3 http://ports.ubuntu.com/ubuntu-ports focal-backports InRelease [108 kB]         
Get:4 http://ports.ubuntu.com/ubuntu-ports focal-security InRelease [114 kB]              
Get:5 http://ports.ubuntu.com/ubuntu-ports focal-updates/main arm64 Packages [1,631 kB]   
Get:6 http://ports.ubuntu.com/ubuntu-ports focal-updates/main Translation-en [395 kB]      
Get:7 http://ports.ubuntu.com/ubuntu-ports focal-updates/main arm64 DEP-11 Metadata [275 kB]                          
Get:8 http://ports.ubuntu.com/ubuntu-ports focal-updates/universe arm64 Packages [938 kB]                               
Get:9 http://ports.ubuntu.com/ubuntu-ports focal-updates/universe Translation-en [234 kB]                             
Get:10 http://ports.ubuntu.com/ubuntu-ports focal-updates/universe arm64 DEP-11 Metadata [407 kB]                              
Get:11 http://ports.ubuntu.com/ubuntu-ports focal-backports/main arm64 DEP-11 Metadata [7,992 B]                                 
Get:12 http://ports.ubuntu.com/ubuntu-ports focal-backports/universe arm64 DEP-11 Metadata [30.5 kB]                       
Get:13 http://ports.ubuntu.com/ubuntu-ports focal-security/main arm64 DEP-11 Metadata [40.7 kB]                           
Get:14 http://ports.ubuntu.com/ubuntu-ports focal-security/universe arm64 DEP-11 Metadata [93.6 kB]                        
Err:15 https://repo.download.nvidia.com/jetson/common r35.1 InRelease                                                      
  Could not connect to repo.download.nvidia.com:443 (23.199.62.69). - connect (113: No route to host) Could not connect to repo.download.nvidia.com:443 (23.199.62.73). - connect (113: No route to host)
Err:16 https://repo.download.nvidia.com/jetson/t194 r35.1 InRelease                       
  Unable to connect to repo.download.nvidia.com:https:
Fetched 4,388 kB in 3s (1,263 kB/s)                                                       
Reading package lists... Done
Building dependency tree       
Reading state information... Done
1 package can be upgraded. Run 'apt list --upgradable' to see it.
W: Failed to fetch https://repo.download.nvidia.com/jetson/common/dists/r35.1/InRelease  Could not connect to repo.download.nvidia.com:443 (23.199.62.69). - connect (113: No route to host) Could not connect to repo.download.nvidia.com:443 (23.199.62.73). - connect (113: No route to host)
W: Failed to fetch https://repo.download.nvidia.com/jetson/t194/dists/r35.1/InRelease  Unable to connect to repo.download.nvidia.com:https:
W: Some index files failed to download. They have been ignored, or old ones used instead.

It works fine on the other two Jetson Xavier NXs. I have checked the DNS configuration on both… and they look identical. Running a host repo.download.nvidia.com on the all three units produces the following output:

repo.download.nvidia.com is an alias for repo.download.nvidia.com.edgekey.net.
repo.download.nvidia.com.edgekey.net is an alias for e28142.d.akamaiedge.net.
e28142.d.akamaiedge.net has address 23.199.62.69
e28142.d.akamaiedge.net has address 23.199.62.73

Interestingly, running a ping on both of those IPs works from the two working units but not from the third. Running a ping to something like google.com works on all three. All three units are connected to the same network switch. I can’t figure out why two work and the third does not… but only to the Nvidia repo… not the other ones.

Any advice or thoughts on what I should check next?

Hi,
Could you try to install SDK Components through SDKManager? See if this method works.

If above method does not work, we would suggest re-flash the system and then try again.

For the three Xavier units above, I did use SDK Manager to set them up. I can tell you, it’s a painful process. One of the units setup just fine. The second unit would not install properly until I swapped the NVMe drive for a new one. The third unit has this APT problem but everything else works. I would rather not have to re-flash the unit but if that’s what I have to do to make it work… then I guess I don’t have much choice. My problem is I can’t see what the flashing has to do with why it’s not connecting to one repo … especially when the other repos work just fine.

I have 3 more of these boxes to setup… it’s a way more difficult task than it should be. Sometimes it works and sometimes it doesn’t… but this seems to be the way things go with anything CUDA related… LOL

UPDATE: Oh, I see what you mean… update it with SDK manager rather than the repo. That actually might fix the problem as well. Will give that a shot.

@DaneLLL … thanks for the advice… I went ahead and re-flashed the box. Since it already had a running version of the JetPack, I did not have to put it into force recover. You still have to drag out a monitor and keyboard to figure out the IP address after flashing, but it worked!

I have to reload my software setup, but the new flash does not have the APT issue. Very strange. Everything on the box is exactly the same (at least as far as I can see) but apt now works for the Jetson repo. Not sure what the problem was.

Anyway, on to the next effort :)

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.