Installation issue - JetPack

Hi people,
I am trying to install the JetPack3.2 on my TX2, but I got big problems.

The installation runs well on my computer, but as soon as I want to install it on my TX2 I got a lot of different issues:

1- I have connected my TX2 and the host pc (running Ubuntu 14.04.05 Trusty) to a switch which is connected to internet, though it can’t connect to a port.ubuntu.com (IP: 91.189.88.150).
Then it tells me it can’t continue because of another ‘process from apt-get is currently running, make sure these packages are installed before continuing…’ with a list of package.

2- I changed the server of connection for the package from France to Main

3- I added new repository

Still with all these changes some package (libvisionworks-samples for instance) remain totally “unable to locate”

I am also wondering if it’s normal that I don’t have internet on the TX2 even if I’m connected to the switch…sometimes the host pc struggles to reach the TX2 so I set the usb(0) connection which improved things but not the unfindable package.

I feel a little desperate and I don’t know what to do…

Thanks in advance for your help

Vincent

JetPack3.2 now has a bug when it downloads the repository.json file (this file specifies URLs of where to download components). Basically 3.2 was superseded by 3.2.1, and I think the deprecation of 3.2 was not completed correctly. Try 3.2.1 instead.

Although the USB ethernet is a good idea I don’t know if it has actually been fully set up and tested yet. I would stick to eth0.

If you do still have problems, then give more information on whether your router is an appliance or a Linux PC. Also add information on what the Jetson sees as its “route” command output, plus “ifconfig” output.

Hi linuxdev,

sorry for the misspell but I actually tried the JetPack3.2.1, I even uninstalled and tried with the 3.1 and I have the same issue…

My router is a netgear device (FS608 v3). All the lights are up no problems. But I can’t get internet on the Jetson which is weird because it is actually connected to the switch (and the Jetson detects the connection).

The weird things for me is when the host PC is looking for the target through eth0 it finds it at a certain IP address, but the critical moment is when it has to download from port.ubuntu.com. It doesn’t reach it and then the installation of the component (CUDA,Visionworks etc…) doesn’t work for a lack of package…

For the output of route and ifconfig command see the 2 blocks below (I did it for the jetson and the host pc for comparison).

JETSON

nvidia@tegra-ubuntu:~$ route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         v196-rtr-doua2. 0.0.0.0         UG    100    0        0 eth0
134.214.196.0   *               255.255.254.0   U     100    0        0 eth0
link-local      *               255.255.0.0     U     1000   0        0 eth0
dhcp1.rocad.fr  v196-rtr-doua2. 255.255.255.255 UGH   100    0        0 eth0

nvidia@tegra-ubuntu:~$ ifconfig
eth0      Link encap:Ethernet  HWaddr 00:04:4b:a5:57:a5  
          inet addr:134.214.197.250  Bcast:134.214.197.255  Mask:255.255.254.0
          inet6 addr: fe80::b102:ec3c:7f9f:d4e/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:4753 errors:0 dropped:0 overruns:0 frame:0
          TX packets:871 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:498138 (498.1 KB)  TX bytes:85084 (85.0 KB)
          Interrupt:42 

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:65536  Metric:1
          RX packets:346 errors:0 dropped:0 overruns:0 frame:0
          TX packets:346 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1 
          RX bytes:41128 (41.1 KB)  TX bytes:41128 (41.1 KB)

wlan0     Link encap:Ethernet  HWaddr 00:04:4b:a5:57:a3  
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

nvidia@tegra-ubuntu:~$ 
/code]


<b><u>HOST</u></b>
[code]vdufour@vdufour-Ampere303R036:~$ route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         v196-rtr-doua2. 0.0.0.0         UG    0      0        0 eth0
134.214.196.0   *               255.255.254.0   U     1      0        0 eth0
vdufour@vdufour-Ampere303R036:~$ ifconfig
eth0      Link encap:Ethernet  HWaddr d8:9e:f3:17:25:47  
          inet addr:134.214.197.130  Bcast:134.214.197.255  Mask:255.255.254.0
          inet6 addr: fe80::da9e:f3ff:fe17:2547/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:15265 errors:0 dropped:0 overruns:0 frame:0
          TX packets:7435 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:13933599 (13.9 MB)  TX bytes:973912 (973.9 KB)
          Interrupt:16 Memory:f7100000-f7120000 

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:65536  Metric:1
          RX packets:804 errors:0 dropped:0 overruns:0 frame:0
          TX packets:804 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1 
          RX bytes:146355 (146.3 KB)  TX bytes:146355 (146.3 KB)

From what I can see they are on the same network…

Thank you

The kind of error I get during installation, sometimes I can find some of the package and manage to install it manually but sometimes there isn’t any “candidate”…

nvidia@tegra-ubuntu:~$ sudo apt-get install cuda-toolkit-9-0 
[sudo] password for nvidia: 
Sorry, try again.
[sudo] password for nvidia: 
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
 cuda-toolkit-9-0 : Depends: cuda-samples-9-0 (>= 9.0.252) but it is not going to be installed
                    Depends: cuda-documentation-9-0 (>= 9.0.252) but it is not going to be installed
E: Unable to correct problems, you have held broken packages.
nvidia@tegra-ubuntu:~$ sudo apt-get install cuda-samples-9-0 
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
 cuda-samples-9-0 : Depends: libxmu-dev but it is not installable
                    Depends: libxi-dev but it is not installable
E: Unable to correct problems, you have held broken packages.
nvidia@tegra-ubuntu:~$ sudo apt-get install libxmu
libxmu6   libxmuu1  
nvidia@tegra-ubuntu:~$ sudo apt-get install libxmu-dev
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Package libxmu-dev is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

E: Package 'libxmu-dev' has no installation candidate
nvidia@tegra-ubuntu:~$ sudo apt-get install libxi-dev
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Package libxi-dev is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source

E: Package 'libxi-dev' has no installation candidate

Here is the ifconfig and route command after flashing and during the process of installation

nvidia@tegra-ubuntu:~$ ifconfig
docker0   Link encap:Ethernet  HWaddr 02:42:08:a3:28:24  
          inet addr:172.17.0.1  Bcast:0.0.0.0  Mask:255.255.0.0
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0 
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B)

eth0      Link encap:Ethernet  HWaddr 00:04:4b:a5:57:a5  
          inet addr:134.214.197.250  Bcast:134.214.197.255  Mask:255.255.254.0
          inet6 addr: fe80::f94f:edda:bf79:e6da/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:1001357 errors:0 dropped:0 overruns:0 frame:0
          TX packets:482506 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:1501064925 (1.5 GB)  TX bytes:51359151 (51.3 MB)
          Interrupt:42 

l4tbr0    Link encap:Ethernet  HWaddr 02:79:85:f1:80:3c  
          inet addr:192.168.55.1  Bcast:192.168.55.255  Mask:255.255.255.0
          inet6 addr: fe80::8de:25ff:fea7:d7d1/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:914 errors:0 dropped:0 overruns:0 frame:0
          TX packets:66 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:171501 (171.5 KB)  TX bytes:6479 (6.4 KB)

lo        Link encap:Local Loopback  
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:65536  Metric:1
          RX packets:250 errors:0 dropped:0 overruns:0 frame:0
          TX packets:250 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1 
          RX bytes:21212 (21.2 KB)  TX bytes:21212 (21.2 KB)

usb0      Link encap:Ethernet  HWaddr 02:79:85:f1:80:3c  
          inet6 addr: fe80::79:85ff:fef1:803c/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:461 errors:461 dropped:0 overruns:0 frame:461
          TX packets:464 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:86712 (86.7 KB)  TX bytes:93920 (93.9 KB)

usb1      Link encap:Ethernet  HWaddr 6e:92:0d:bf:cb:bc  
          inet6 addr: fe80::6c92:dff:febf:cbbc/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:453 errors:0 dropped:0 overruns:0 frame:0
          TX packets:462 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:84789 (84.7 KB)  TX bytes:73557 (73.5 KB)

wlan0     Link encap:Ethernet  HWaddr 00:04:4b:a5:57:a3  
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000 
          RX bytes:0 (0.0 B)  TX bytes:0 (0.0 B) 

nvidia@tegra-ubuntu:~$ route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         v196-rtr-doua2. 0.0.0.0         UG    100    0        0 eth0
134.214.196.0   *               255.255.254.0   U     100    0        0 eth0
link-local      *               255.255.0.0     U     1000   0        0 docker0
dhcp2.rocad.fr  v196-rtr-doua2. 255.255.255.255 UGH   100    0        0 eth0
172.17.0.0      *               255.255.0.0     U     0      0        0 docker0
192.168.55.0    *               255.255.255.0   U     0      0        0 l4tbr0

Generally at each installation of a component (CUDA,Visionworks,TensorRT…) it manages to do some of the installation but there are always packages it can’t get from the port.ubuntu.com and so ask me to install it manually.
From what i can see the libopencv-dev is always problematic, just as the cuda-samples, libvisionworks-samples and other following for each components.

Just like here it can’t even locate the package…

nvidia@tegra-ubuntu:~$ sudo apt-get install libvisionwork-tracking
Reading package lists... Done
Building dependency tree       
Reading state information... Done
E: Unable to locate package libvisionwork-tracking
nvidia@tegra-ubuntu:~$ sudo apt-get install libvisionwork-tracking-dev
Reading package lists... Done
Building dependency tree       
Reading state information... Done
E: Unable to locate package libvisionwork-tracking-dev

So at the end there is maybe half of the whole Jetpack correctly installed…

Do be sure that if you use 3.2.1 after some other version that the directory you run it from is clean. I think there are cases of old JetPack versions (or just different versions) not behaving well together.

The Netgear FS608v3 is a fairly simple and ordinary 100Mbit/s switch and should be fine for connection. However, the switch isn’t a router and does not hand out IP addresses, it simply passes a DHCP request/response through.

So the next question is how does the Netgear FS608v3 connect to your router? What is your router (perhaps it is a cable modem or some other modem which also acts as a router)? This is how the DHCP will query for IP address. Both the Jetson and the host PC probably go to the switch, but then the switch still needs a router to the internet. During JetPack operation it’ll ask whether you want to use your PC as a router, or if you have an actual router appliance…which is basically what I’m asking because this says who will reply to a DHCP IP address request. How did you answer when JetPack wanted to know if you will use the PC as a router?

Is your router used only by you? I’m just wondering if it might be shared, e.g., in a college dormitory or a business. If there are other Jetsons on the inside of the router, then it may be someone else’s Jetson replying.

This says your Jetson did get a valid address:

inet addr:134.214.197.250

On the other hand, I see a docker device. If there is any use of a docker container to install software, then you’ve probably found the culprit. Can you give more details about whether docker is involved in your flash? Routing would probably be quite messed up in that case with not all traffic getting through, but with some cases of partial success.

For the directory I always clean it or even delete it to recreate it so should be fine.

Sorry my switch is a Netgear and I connect it to a RJ45 port which is directly in the wall. My network is a public institution network so I just have to connect a wire from a plug in the wall and that’s it.

I can ask if other Jetson are connected but as far as I know my department is the only one.

When the Jetpack asks I said that I have the first layout meaning the TX2 and my host are connected separately to the router (through the switch). Though I could totally try with selecting the second solution…

What is the role of this docker ? I don’t know anything about it. How does it disturb my installation ?
How can I obtain the information you want concerning the docker involvement ?
Can I just inactivate the docker interface during the process ?

Also I still have this issue concerning internet connection: I can’t go to google even if the TX2 is connected…

Thank you a lot for taking the time to help me

There may be an issue with the router since you don’t control it. Public institutions may limit allowed MAC addresses, so on (and you need to immediately change all account passwords after installing since the password and login names are known by the whole world…two Jetson owners could accidentally log in to the other’s Jetson and not even know it). Normally I would tell you to use the PC as router instead of the actual router, and this might work, but when I do that I have the Jetson on a private second network card. Perhaps it would work, perhaps not…it is worth trying. So, I suggest you start a new JetPack install and let it use PC as router…if it fails, just go back to the previous configuration.

You might also want to ask the admin of the router if there is any chance it was firewalled or denied. When communicating with the admin you can tell him the ethernet MAC of your Jetson is “00:04:4b:a5:57:a5”.

In your ifconfig output you will see a docker interface. This means you have a virtualization container going on. If any virtualization is being used for the install, then you will be breaking the chain of network routing unless you’ve specifically configured it correctly. Have you set up any kind of virtualization or VM software? Any interaction with this would likely break things.

Ok I will change the password after flashing that’s a good point.

I can try using the host as a router but I am not sure there is a second network card in my host (it’s Dell Precision Tower 3620, core i7)… I will check but even doing this do I have to change my set up (or do I let the Jetson and host connected to the switch) ?

Okay I will ask thank for the precision.

Concerning the docker, I had read that VM wasn’t a good idea so I definitely didn’t do it.
Is there any way to delete this docker ?

Look at the screen-shot to have an idea of how the different errors I get appear and go on.
(Sorry but not possible to copy from this window so I took picture…)

The only things I did was to install Ubuntu on my computer on a separate hard-drive using a flashed-USB, then connect the switch (host + jetson) and follow the install guide…

The problem with using the PC as router in your case is that the network will send out a request which both the router and the PC see…both will try to answer. So it probably isn’t a good idea to use host PC as router when you can’t control the router appliance itself. If you do choose this it is just a button press in JetPack when it asks which method to use.

I don’t know what you have using docker. It is quite possible this is confusing networking, but I have no way to advise on it. Someone who knows docker better might be able to tell you how to trace whatever is actually creating or using the docker network interface.

There is a good chance that docker setup is altering things, though it might have nothing to do with it. I see an address your host is failing to reach is IPv6 format (usually it is IPv4 format…sometimes a lack of IPv6 features can interfere) “2001:67c:1360:8001:22” for “ports.ubuntu.com”. Can you provide what you see from these commands without rebooting or configuring anything between commands:

host ports.ubuntu.com
route
traceroute 2001:67c:1360:8001::22

Keep in mind that it isn’t always a problem on the PC when there is a network issue, e.g., if you request a package which doesn’t exist you’ll see an error, but there are often multiple mirror servers and it would fail over.

The part which is more likely to be of interest is where the picture shows:

Error: CUDA cannot be installed on device. ...
...
cuda-toolkit-9-0 libgomp1 libfreeimage-dev libopenmpi-dev openmpi-bin
After these packages are installed on device, press Enter key to continue

If you look at your JetPack directory you should see a “jetpack_download/” subdirectory. Within this what do you see from:

ls cuda-repo*

If you see that file, and it is “arm64.deb”, then this file is for install on the Jetson. You could try to manually copy this to the Jetson and run:

sudo dpkg -i
# ...if that succeeded, then also:
sudo apt-get install -f

If you see an “amd64.deb” file, then this is for your host instead, but that’s a different topic.

Note that the repo package adds a set of “.deb” files to a subdirectory of “/var” and then tells apt to treat this the same as a repo somewhere on the internet. Those files become available for automatic dependency resolving when you use apt. Without this repo you won’t be able to succeed. So probably the first thing to do is figure out if this repo .deb can be installed, or figure out why it is failing. After that branch off to fix other issues.

Any time you are getting any message about there already be an apt or dpkg operation running and that it can’t do this at the same time see what the result is from:

sudo lsof /var/lib/dpkg/lock

(note whether such error is from trying to install arm64 versus amd64 files…arm64 is Jetson, amd64 is host…you want to check lsof of the lock file on the machine which is failing)

When the device tries to connect to ports.ubuntu, I see the IPv4 address (see the image attached), but then when the errors come in it shows the IPv6…

Here is the result of the commands you asked (on the host).
Is it possible that a firewall prevents my host from connecting ?

vdufour@vdufour-Ampere303R036:~$ host ports.ubuntu.com
ports.ubuntu.com has address 91.189.88.150
ports.ubuntu.com has IPv6 address 2001:67c:1360:8001::22
vdufour@vdufour-Ampere303R036:~$ route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         v196-rtr-doua2. 0.0.0.0         UG    0      0        0 eth0
134.214.196.0   *               255.255.254.0   U     1      0        0 eth0
vdufour@vdufour-Ampere303R036:~$ traceroute 2001:67c:1360:8001::22
traceroute to 2001:67c:1360:8001::22 (2001:67c:1360:8001::22), 30 hops max, 80 byte packets
connect: Network is unreachable
vdufour@vdufour-Ampere303R036:~$

I indeed see the 2 files in my jetpack/download directory, I copied it and execute the command you told me to and here is the result:

vidia@tegra-ubuntu:~$ sudo dpkg -i cuda-repo-l4t-9-0-local_9.0.252-1_arm64.deb [sudo] password for nvidia: 
(Reading database ... 149267 files and directories currently installed.)
Preparing to unpack cuda-repo-l4t-9-0-local_9.0.252-1_arm64.deb ...
Unpacking cuda-repo-l4t-9-0-local (9.0.252-1) over (9.0.252-1) ...
Setting up cuda-repo-l4t-9-0-local (9.0.252-1) ...
nvidia@tegra-ubuntu:~$ sudo apt-get install -f
Reading package lists... Done
Building dependency tree       
Reading state information... Done
The following packages were automatically installed and are no longer required:
  apt-clone archdetect-deb dmeventd dmraid dpkg-repack gir1.2-timezonemap-1.0
  gir1.2-xkl-1.0 kpartx kpartx-boot libappstream3 libdebian-installer4
  libdevmapper-event1.02.1 libdmraid1.0.0.rc16 liblockfile-bin liblockfile1
  liblvm2app2.2 liblvm2cmd2.02 libparted-fs-resize0 libreadline5
  lockfile-progs lvm2 os-prober pmount python3-icu python3-pam rdate
  ubiquity-casper ubiquity-ubuntu-artwork
Use 'sudo apt autoremove' to remove them.
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
nvidia@tegra-ubuntu:~$

If I understood well it says that everything is already installed at the right version… and the file was actually already in the cuda-l4t directory on my Jetson but after those operation I still have only 3 files in this directory: cuda-l4t-error.log, cuda-l4.sh and the arm64.deb.

Thanks for the command to look for the lock apt-get.

Hi Linuxdev,

I finally succeed in the installation ! I obtained a full internet access and lauch the installation and everything went well.

Though there is something weird: the directory cuda-L4t disappeared, cudnn and OpenCV are not installed even if the whole installation went well… I don’t get it.

All these directory and files are present on my host, so do you think a basic copy and run would work ?

Assuming you mean the JetPack directory (not “/usr/local”) of the host: Within the JetPack directory where you ran the install, there is an option controlling whether to delete temporary files…this is what determines if those JetPack files get deleted after JetPack completes (I don’t recall the exact name, look for a checkbox while running JetPack…keep in mind you can uncheck flash and just do package ops without flashing). If that is not checked then those files won’t be removed from the JetPack directory, but default is to remove them after finishing.

If you mean files in “/usr/local/” of the Jetson, then old versions should not be removed (though I suppose it is possible JetPack has an option to remove packages…or one install might overwrite old installs with a failure resulting in removal instead of repopulating). So I guess double check what you have checked for install, but can you verify if you mean the “/usr/local/” files of the Jetson/host?

FYI, the cuda “repo” package is the one which starts everything, and if you have this on your host for the arm64 (not amd64), then you can copy this to the Jetson and manually use “dpkg” to install this (amd64 is for the PC host). Once this is installed other packages can be installed, e.g., the actual CUDA in “/usr/local” (the repo makes the packages available…after that you can search “apt search cuda” and see availability).

Alright I followed your advise and just relauched the installation of package without fully flashing the Jetson and it’s good now I have all the libraries installed.

Thank you a lot for your advanced help and time.

Hi again,

I have a redundant issue: OpenCV4Tegra doesn’t get installed on my Jetson.

I tried to follow the first Tutorial of OpenCV in which it shows how to create a simple window with a message in it and here are the errors coming in:

nvidia@tegra-ubuntu:~/OpenCV4Tegra/0-cv-hello/build$ make
Scanning dependencies of target cv_hello
[ 50%] Building CXX object CMakeFiles/cv_hello.dir/hello.cpp.o
/home/nvidia/OpenCV4Tegra/0-cv-hello/hello.cpp: In function ‘int main()’:
/home/nvidia/OpenCV4Tegra/0-cv-hello/hello.cpp:7:5: error: ‘putText’ is not a member of ‘cv’
     cv::putText(img,"Hello OpenCV",
     ^
/home/nvidia/OpenCV4Tegra/0-cv-hello/hello.cpp:8:23: error: invalid use of member function (did you forget the ‘()’ ?)
  cv::Point(10,img.row/2),
                       ^
CMakeFiles/cv_hello.dir/build.make:62: recipe for target 'CMakeFiles/cv_hello.dir/hello.cpp.o' failed
make[2]: *** [CMakeFiles/cv_hello.dir/hello.cpp.o] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/cv_hello.dir/all' failed
make[1]: *** [CMakeFiles/cv_hello.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2
nvidia@tegra-ubuntu:~/OpenCV4Tegra/0-cv-hello/build$

Which is weird cause when I’m using apt search I got these answers:

nvidia@tegra-ubuntu:~$ sudo apt search libopencv
[sudo] password for nvidia: 
Sorting... Done
Full Text Search... Done
libopencv/now 3.3.1 arm64 [installed,local]
  Open Computer Vision Library

libopencv-apps-dev/xenial 1.11.11+ds-1 arm64
  opencv_apps Robot OS package - development files

libopencv-apps0d/xenial 1.11.11+ds-1 arm64
  opencv_apps Robot OS package - runtime files

libopencv-calib3d-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-calib3d

libopencv-calib3d2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision Camera Calibration library

libopencv-contrib-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-contrib

libopencv-contrib2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision contrib library

libopencv-core-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-core

libopencv-core2.4v5/xenial,now 2.4.9.1+dfsg-1.5ubuntu1 arm64 [installed,automatic]
  computer vision core library

libopencv-dev/now 3.3.1 arm64 [installed,local]
  Development files for Open Source Computer Vision Library

libopencv-features2d-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-features2d

libopencv-features2d2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision Feature Detection and Descriptor Extraction library

libopencv-flann-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-flann

libopencv-flann2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision Clustering and Search in Multi-Dimensional spaces library

libopencv-gpu-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-gpu2.4v5

libopencv-gpu2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision GPU library

libopencv-highgui-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-highgui

libopencv-highgui2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision High-level GUI and Media I/O library

libopencv-imgproc-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-imgproc

libopencv-imgproc2.4v5/xenial,now 2.4.9.1+dfsg-1.5ubuntu1 arm64 [installed,automatic]
  computer vision Image Processing library

libopencv-legacy-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-legacy

libopencv-legacy2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision legacy library

libopencv-ml-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-ml

libopencv-ml2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision Machine Learning library

libopencv-objdetect-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-objdetect

libopencv-objdetect2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision Object Detection library

libopencv-ocl-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-ocl2.4v5

libopencv-ocl2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision OpenCL support library

libopencv-photo-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-photo2.4v5

libopencv-photo2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision computational photography library

libopencv-python/now 3.3.1 arm64 [installed,local]
  Python bindings for Open Source Computer Vision Library

libopencv-samples/now 3.3.1 arm64 [installed,local]
  Samples for Open Source Computer Vision Library

libopencv-stitching-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-stitching2.4v5

libopencv-stitching2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision image stitching library

libopencv-superres-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-superres2.4v5

libopencv-superres2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision Super Resolution library

libopencv-ts-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-ts2.4v5

libopencv-ts2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision ts library

libopencv-video-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-video

libopencv-video2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision Video analysis library

libopencv-videostab-dev/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  development files for libopencv-videostab2.4v5

libopencv-videostab2.4v5/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  computer vision video stabilization library

libopencv2.4-java/xenial 2.4.9.1+dfsg-1.5ubuntu1 all
  Java bindings for the computer vision library

libopencv2.4-jni/xenial 2.4.9.1+dfsg-1.5ubuntu1 arm64
  Java jni library for the computer vision library

nvidia@tegra-ubuntu:~$

As you can see some of these are checked as installed… I’m really confused

I tried several time to reinstall the libraries but each time I have these messages:

nvidia@tegra-ubuntu:~/OpenCV4Tegra$ sudo dpkg -i libopencv-dev_3.3.1_t186_arm64.deb 
[sudo] password for nvidia: 
(Reading database ... 167835 files and directories currently installed.)
Preparing to unpack libopencv-dev_3.3.1_t186_arm64.deb ...
Unpacking libopencv-dev (3.3.1) over (3.3.1) ...
Setting up libopencv-dev (3.3.1) ...
nvidia@tegra-ubuntu:~/OpenCV4Tegra$

And then nothing changes… I rebooted already…, the apt-get install -f doesn’t give anything much just saying I have to upgrade…

I also tried to install by clicking directly on the deb file, it opens a window ‘Ubuntu Software’, I click on install, it installs and then still same errors…

Thanks in advance

My understanding is that OpenCV4Tegra is no longer needed because the newer OpenCV includes CUDA support. I don’t know the details of how to transition from using the older OpenCV4Tegra and the newer mainstream OpenCV which has CUDA support though…someone who has used both can probably tell you what is needed. My guess is that your libopencv-dev package did the trick but software somewhere needs a configuration change.

Okay so the tutorials are no longer valid cause OpenCV changed too much.

I’ll try to find documentation on how to use the newest version.

Thank you for support