It seems to me like power on and off have two different behaviours:
if I shutdown jetson xavier nx through the operating system and after that turns off the power, jetson xavier wont start when I turn the power back on. First I have to remove the powerplug from the device and reinsert it again.
if I just turns off the power and in that way are closing jetson xavier it will restart as soon as I turn the power on again.
While number two works it seems always better to turn off the operating system first but it is a bit odd to remove and reinsert the cable to restart it.
What is the proper way of doing this? Or does number two nothing to worry about? Or am I better off install some kind of on off switch or what is the most easy way of handle this fundamental thing?
It only works in these three scenarios:
cutting the powersupply - the jetson starts when it is back on
shutting down from menu - the jetson starts if I remove and reconnect the powerplug to the jetson
shutting down from menu - the jetson starts if I wait 5-10 minutes before I start the electricity again
Related
My problem is as follows.
From time to time, Xorg starts a process with a heavy CPU load (50-100% only for one thread) and animation a little freezes on the screen (for example, letters are displayed with a delay of about 0.5-2 seconds after they are entered using the keyboard or with micro freezes when video playback).
Please help me figure out what this process does and how to fix the problem with the fact that the graphics has a delay due to it.
Below I leave the full display of the command (taken from "top" command):
Xorg -nolisten tcp -background none -seat seat0 vt1 -auth /var/run/sddm/{efd6b83b-791f-4698-bfmc-7a0badf342c7} -noreset -displayfd 17
I'm using Manjaro (KDE) and this problem has been happening for at least a year despite regular updates.
What I managed to do myself:
If I somewhat restart the graphical shell: press Ctrl+Alt+F2 (displays a terminal) and then return to Ctrl+Alt+F1 (graphics), then the system reports: Desktop effects were restartet due to a graphics reset, sometimes Xorg disappears process (not always).
Sometimes Xorg process disappears on its own
There is a possibility that this may have something to do with Firefox, which is always open.
Additional Information:
There were big problems with graphics from the very beginning of using a laptop with Linux. From time to time I have to restart the graphical shell due to various glitches when using the computer for a long time without restarting (sleep only), say 20 days (for example, some pop-ups are displayed in black with a 50% probability)
Also sometimes plasmashell takes gigabytes of RAM, which leads to a freeze, and is treated only with "killall plasmashell" and restart
Computer: Lenovo Legion Y540
Video card: NVIDIA GeForce GTX 1650 4GB GDDR5
OS: 5.10.161-1-MANJARO (KDE)
I am an intern in charge of improving the gazebo simulation for my internship. Currently, I managed to successfully drive my virtual buggy in the gazebo world smoothly as shown in the video link.
https://youtu.be/uS5PXshZZTU
Right now, I am tasked by my supervisor to connect my virtual buggy's movements to the movement of the real buggy that we have, so for example, when the real buggy moves 1 metre, the virtual buggy also moves 1 metre in its virtual world. I was told to use ROS Bridge Server to do this task but I do not have an idea on how to start this task at all. So I would appreciate it if someone who is knowledgable in this area help me out.
Mostly, it is really simple to change from simulation to real world application. The most ROS driver for robots have the roslaunch flag simulation and setting it on false causes ROS to launch the real controllers to.
I do not know if you created your gazebo package from scratch but if not there must be a possibility to launch the launch file with a sim:=false flag. Look into thr rviz support and in the demo.launch file.
I've taken over the maintenance job of a device driver from another guy recently.
The driver works with a pci-e board. when the system starts up, the driver will allocate a 128Mb memory, and then set some registers on the board to pass the information about the allocated memory. Later when the board is working, it will write some data to the memory and some other application will access those data through the driver.
The board is powered by a standalone adapter (not powered by pci-e slot, it's a demo or development board). So when sometimes the fpga program in the board goes wrong, we will re-powered on the board (this is fast) and restarted the pc also (this is slow, otherwise the board will know nothing about the driver-allocated memory).
Here comes my problem: is it possbile for driver to know the board has been re-powered on? if it can detect the event and do something like those done when the system starts up, it will save us much time.
I haven't got much knowledge about driver development before, it would be helpful if there is any not-too-complicated tutorial/article for this kind of job. Or maybe I must find some thick books to learn from scratch?
Hope I've made myself understood and any suggestions would be greatly appreciated:-)
Check Supporting PnP and Power Management in Function Drivers might help you.
I just read an article about WOL vs IPMI, it says "...used WOL to save millions of dollars by powering off idle desktops"(almost end at last paragraph but one), does it mean that WOL can be used to power off a desktop?
If so, how?
p.s. Does mainstream server support IPMI?
No, Wake On LAN lets you turn on PCs, not turn them off. The idea is that PCs need to be on in the middle of the night so they can get automatic updates; rather than leaving them on all day and all night just to get occassional updates, you can turn them off at the end of the day and use WOL to turn them on only when you need to update them.
I am writing device software for a PC and for that, I want the PC to be usable as a device. When power is supplied, it should switch on without requiring to press the power button. There are power options in BIOS settings but it starts the PC only when its uncleanly shutdown. The other concern I have is how would unclean shutdown affect the hard disk, filesystem and the OS (XP or Linux).
What you need is another PC and one of these devices attached to it.
http://www.relaypros.com/mm5/merchant.mvc?Screen=CTGY&Store_Code=NCD&Category_Code=RS-232_Relay_Boards&gclid=CMna8_yOo5wCFQxM5QodWjoflQ
What you do is send this some RS232 commands for a quick closure on one of the relays. The relay is connected to the Power On pins of the computer you want to control.
You possibly could find another relay contact closure for AC current that allows you to close a relay when AC is flowing, but you would only want to for a brief second.
Unsafe shut downs can be quite detrimental depending where the filesystem state is in. It would be quite hard on the hardware too.
There is also the alternative of booting from the network device. A quick search led to some information on wikipedia. Also, there is something related called preboot execution environment which seem to be something like what you are looking for.
Some software options - these aren't exactly what you asked for, but they might help
Mac OS X: In the energy saver control pane's options tab, select "Restart automatically after a power failue. shutdown -hu now should then bring the system down but give you 5 minutes to remove power to simulate a dirty shutdown, and have the computer reboot automatically when power is restored. It's a slightly dirty shutdown anyway, I think. (ie, it doesn't log you off first)
Windows:
I don't have a windows machine so I can't try this, but you used to be able to tell windows not to power down the computer when you select shut down, but rather to put it in a safe state and display "It is now safe to turn off your computer". Perhaps you could then remove the power and have the bios believe it was a non-clean shutdown, and turn the machine on again when power is restored. There are some instructions on how to do this in Windows Server 2003 at the bottom of this microsoft help document. This forum discussion seems to suggest it might work on XP.
Linux: Not sure about this one, but maybe this website can help.
I haven't tried any of these, so no guarantees that they'll work or work safely.