controller spawner error/ur_hardware_interface died #370
Labels
close if no update
Solution proposed
A solution has been proposed inside the issue. Waiting for user feedback.
Comments
|
I meet the same problem too. Two days ago, I can successfully start my ur5, but it didn't work since yesterday. I did the rebuild too, but it didn't work. I am on Ubuntu 18.04 and ROS Melodic too. But sometime I can successfully connect, the probability is almost 5%. If this problem happen to most of computer suddenly, I guess it may cause by update of some package. I hope anyone can solve this problem, because my work should be finish in 3 days. Btw, I can successfully connect to the ur5 by ur_bringup everytime(Trough it isn't support Melodic? ), but I found the tf tree is broken, so I think it isn't a good way to solve this problem. |
|
Uh, that sounds interesting! We did indeed a change in the library regarding tcp_server implementation. For this to fully understand, I would kindly ask you to help us debugging this:
|
|
Hey, thanks for your answer. But I can't do these right now because I have no access to the robots until monday, so I will have to try these next week. I hope this isn't an issue. I might look into my update logs what exactly was updated, maybe this might help narrowing it down. |
oh yeah, the client-library hasn't been synced to the main ROS repository, yet, so that's probably not the source of these problems. (Given that you don't build the library from source inside your workspace, already). |
|
SO my update log from the 5th of may: I included everything ros-related, my assumption would be that it's ros_control/ros_controllers? What I found googling around is that some "ABI breaking changes were pushed" and it should be all done updating every package/rebuilding the workspace (as mentioned above). Also the error seems controller related (sadly, can't post the full logs right now). But I am already up to date I guess and rebuild the workspace more than I want to admit. I also think it's kind of weird that I had so many updates open, even though installing ROS Melodic not long ago (maybe 2-3 months ago). Is there a reason for such a thing? Anyway, I hope this does help a bit! |
Hey, fmauch. The following is my full output of I uninstall my |
This seems as if the driver came up, correctly, but then crashes without any further output. Could you please try to start the robot inside gdb, using the Then, you will end up in a gdb prompt similar to this: There, please enter The output of that should show us where that actually crashes. |
Hey, fmauch. Thanks for your answer. The following is some of my output of |
|
Hm, that looks different to your previous output, as we see the output |
|
Alrighty, I am back on the robots and this is my terminal: I also tried a few times to get the right output out for comparison: The shutdown request in between can be ignored, everything still works fine and problems occurred after updates! I hope this does help? Edit: the shutdown request was from another rviz which I forgot to close before connecting to the robots (oops, sorry). It can be completely ignored. Edit part 2: (sorry for those) I did the debug:=true method which wouldn't work because of ROS killing processes before I could enter bt. Anyway: This is what happens when I use debug:=true but the controllers actually load: |
|
So, today I ran again sudo apt upgrade and got a huge amount of packages to upgrade (practically everything). Well, things got way worse now. I can't get the driver to start: And sometimes somthing like this happens: |
|
After updating, have you cleaned your workspace and then rebuilt it? |
|
I just did that and now the driver won't build (among other things) giving me this error: |
|
You'll probably need to But it's a bit difficult to read, as you show lots of warnings for packages not the UR driver. |
|
Sorry about that, I removed some packages and this is the error I get when I catkin_make_isolated: ur_calibration inside the driver seems to be the problem, I assume? |
|
You seem to have a Did you update that? |
|
I used the commands fmauch wrote in his earlier post from 6 days ago not sure if that is updated though. |
|
I'm not sure which commands you're referring to. Please always link to instructions you've followed. I'm not near a machine where I can check what's going on, but if there are changes in the client library the UR driver is expecting and you don't have those in your local checkout, things won't compile. |
|
Oh sorry, #370 (comment) this comment is what I meant. I will keep looking, thanks. |
|
The errors showing up in your build output have been addressed inside #330. Updating both, the client library and the driver to the latest master version using |
|
Yep, thanks. I just saw that in the git commit and changed it by hand (since it's just syntax). Now it built! Buuuut, still the same error as earlier. Might roll back or even reinstall everything at this point :/ thanks for the help though. |
The So, which output are you referring to? |
|
Sorry for the late answer, this is the output I am referring to: #370 (comment) |
|
So, the relevant gdb output seems to be Just for clarification: Did you clean your environment (delete your build and devel folder of your workspace) after updating the ROS packages? I would just like to rule out any ABI incompatibilities... If not, this could point to a problem inside our scaled trajectory controller implementation. |
|
Did that a few times (more than I want to admit lol). Something new I noticed now that I got to build my workspace again is that message/warning once a connection goes through and the driver actually starts: The warning is getting published god knows how many times a second but doesn't disturb (other than spam). |
|
That should not necessarily be related. Do you use a custom urscript file? Because, this has been addressed in 13db13e. If you don't use your own script file, that might be a clue that we're still not using the latest master here. |
|
Nope, no custom urscript file. So it seems I am not on master... Should I just clone the repo from master? |
|
It looks like you are starting the driver from a more complex launchfile. Maybe let's break it down by just using the ur10_bringup.launch file. Then, we'll see the driver's parametrization in the output. Additionally, please go to your package clone ( |
|
I'll do these (again) next week on monday when I'm back in office (sorry, am an intern, don't have the time to access the robots). Will report back when I tried it with just one robot. Anyway, I wouldn't call them that complicated, just one bringup calling ur_common etc for each robot (every robot has his own controllers.yaml and own launch file, probably one hell of an ineffiecent way to do that). Thanks for your time though and sorry for the wait. |
Sorry, this is my full output of
Is my problem different from UmutUz's? Should I open a new issue? |
|
Last week I had the amost the same question, too. I opened the workspace I had been done successfully but the error said:
|
|
So, I managed to get back to the robots(finally!). Sorry for the very late answer: I cloned the repo again and and it's still not working it's the same errors. Since it worked out for @qwq321qwq I might try updating urcap too (if I am allowed to lol). |
I got error again but if i changed the position it could work. So I write to ask in another issue. |
|
Well, I am not allowed to update urcap. I am stuck I guess.. |
I get the same warning message as you. Additionally, when I click play on my External Control urcap, the URSim gets stuck on a 'Starting program... Please wait..' prompt. I still get the message saying that the robot is ready to receive control commands, though. |
yes,it showed the error but could run the controller still |
Tried running a simple action client to '/scaled_pos_joint_traj_controller/follow_joint_trajectory', but it doesn't seem to work. Same problem as in issue #305, but I have the correct port 5002 set up as well as inactive firewall on both sides. |
|
@UmutUz I was facing similar issues to you. |
Update: turns out I made a very silly mistake. Was using port 5002 instead of 50002. |
I switched from my own laptop back to a jetson nano given from work which doesnt have that problem (for now). Thanks for the tip tho! If this happens again I will try your suggestions out :) |
stefanscherzinger
added
the
Solution proposed
|
This issue has not been updated for a long time. If no further updates are added, this will be closed automatically. Comment on the issue to prevent automatic closing. |
|
This issue has been closed due to inactivity. Feel free to comment or reopen if this is still relevant. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
For a few days now, I have been getting the same problem over and over.
Usually, I would enter my command roslaunch ur_robot_driver ur10_bringup.launch [reverse_port:=REVERSE_PORT] and both my UR10 robots would start with no problem.
Now, it's hit or miss. More often than not, the controllers crash and I get the text on top.
Looking around, I should either 1)update all my ROS packages 2)rebuild my workspace 3)both.
I did both (in fact, I did the rebuild part a lot the last few weeks) and my assumption would be that updating caused these problems. Before I updated all my packages it was working fine. This is the command I ran to update and upgrade
sudo apt update && sudo apt upgradeI am on Ubuntu 18.04 and ROS Melodic. I did google and tried again the stuff mentioned above. I don't know why it's still hit or miss(more often miss than hit sadly)
The text was updated successfully, but these errors were encountered: