You are not logged in.
Pages: 1
Hi all, hope someone can help me with the following:
I am running Arch Linux 4.4.0-104-generic (x86_64) on my laptop. I am updating once a week; about 2 weeks ago, right after updating, my computer failed to start after rebooting. Unfortunately, this problem proved to be inconsistent, meaning my computer sometimes does start at once, sometimes starts after one or more reboots, and sometime does not start at all (I have a fall-back system on another partition, with which I have not seen this problem). Also the Arch fallback system acts similar.
I have done several things to investigate the problem, but do not seem to get any closer; I have added ¨systemd.debug-shell=1 systemd.log_level=debug" to the kernel line in Grub, and deleted "quiet" as to get more output during boot, and hopefully to be able to start a terminal to further investigate. Although this did work sometimes, it does not most of the time, so I am beginning to believe that the problem occurs before systemd is started.
When I start my computer, I get to the grub menu, and from there select and start Arch; the menu disappears, and the last two "echo" lines of the boot sequence as in Grub appear on my (further black) screen ("booting Linux", "loading RAM disk"), followed by a (not blinking) cursor. There is for about a minute or so disk activity; I also see after ~10 seconds the webcam led flash, as if it is being initialized. This is where it ends, and where after I reboot and hope for the best. As mentioned, this sometimes works, sometimes not. I finally added "nomodeset" to the Kernel boot line, but again to no avail.
I am kind of lost here, and hope someone can help me out.
Some computer info:
Intel(R) Pentium(R) Dual CPU T2390 @ 1.86GHz, 3GB Ram, Medion Akoya. I am using the NVidia 340 type graphics driver, which I suspected at first since it was updated right before the system started to act up. It was since upgraded again to 340.104-16, but the problem is still there.
Offline
Arch Linux 4.4.0-104-generic
What kernel is that?
Offline
What kernel is that?
It would appear to be a Debian/Ubuntu offering, however i'm sure if OP was dual booting he would have mentioned it. So as per usual .. ¯\_(ツ)_/¯
My guess is bootloader is pointing to a non existent or incompatible kernel.
Last edited by cirrus (2018-01-02 21:17:21)
Ancestoral Clan https://cirrus.freevar.com/mclean.html
Offline
Ooops, thanks for replying!
The listed kernel is indeed the one of my fall back system (as mentioned), and indeed this is an Ubuntu derivative...
I believe the correct one is 4.14.8-1!
If the bootloader is pointing to a wrong point,wouldn't the system refuse to load at all?
Last edited by luigi01 (2018-01-02 21:42:51)
Offline
Please post the output of the following commands:
uname -a
pacman -Q nvidiachroot into the system if you can't boot it.
Last edited by 2ManyDogs (2018-01-02 21:47:38)
Offline
https://wiki.archlinux.org/index.php/Ge … t_problems
https://wiki.archlinux.org/index.php/GR … ramebuffer
In doubt try to boot "nomodeset"
Also see whether the failing boot already write into a journal and sync to disk.
What's your login manager?
Offline
Tried to start Arch last night 5 times, no luck: did not start, no output to screen, no system terminal accessible.
This morning however all went well - following commands were run in a running Arch system:
uname -a:
Linux archlinux 4.14.8-1-ARCH #1 SMP PREEMPT Wed Dec 20 21:27:44 UTC 2017 x86_64 GNU/Linux
pacman -Q nvidia:
error: package 'nvidia' was not found
output of pacman -Qs nvidia:
local/libvdpau 1.1.1+3+ga21bf7a-1
Nvidia VDPAU library
local/nvidia-340xx 340.104-16
NVIDIA drivers for linux, 340xx legacy branch
local/nvidia-340xx-utils 340.104-1
NVIDIA drivers utilities
I tried to add "nomodeset" (at a time the system did not boot), no change;
my login manager is LXDM
Also checked journalctl, but I believe there were no entries from the no-boot-times.
Will add some suggestions from the "general trouble shooting", "boot problems" wiki.
At one time, when the system did not freeze right away, I did manage to start the systemd-shell (ctrl-alt-F9),
ran journalctl, and saw an abbreviated error message "Nvidia probe routine failed.....", "None of the Nvidia adapters were....";
the terminal size was standard (80x25 ??), making me set a higher resolution (grub-gfxmode=1440x900 as checked using grub vbeinfo) -
the menu looked fine, but no more output was generated during boot - I just changed this back to 'auto' (for the next boot).
What really puzzles me is that booting is fine one time, and impossible the next time, without any change in between.
Offline
The graphical.target won't work w/ nomodeset - you'll have to boot the multi-user.target
Offline
Well, still no improvements: at random my pc freezes right after echoing the last 2 lines in my Grub-menu .....
echo 'Loading Linux linux ...'
echo 'Loading initial ramdisk ...'
....followed by a frozen cursor on the following line.
So, nothing is started, no systemd, no logs, nothing to check. I checked the UUID against the blkid from the corresponding partition, no difference found.
The only thing I can (?) do is to reboot and hope things will run at the next boot.
I am out of ideas, any help is extremely appreciated!
Offline
Let's determine the basic status of the system. Can you ping it from another machine when this happens?
If not, did you ever get this with the lts kernel?
What if you use the debug initramfs?
Finally pelase post the output of "cat /proc/cmdline" (just a hunch)
Offline
Maybe earlyprintk kernel parameter might be helpful here.
https://raw.githubusercontent.com/torva … meters.txt
sys2064
Offline
After several successful boot ups, today things went bad again - needed 4 boots before the system decided to run.
What I noticed: as soon as Grub times out (5 secs) the screen is cleared, and the last two 'echo' lines out of Grub are printed to the screen; as mentioned, the disk led flashes for some time, and I notice the webcam led flashing once as if the webcam was initialized, but the screen continues to be frozen.
When the system at the fourth try eventually did start, the first line (after the Grub messages are cleared) on the screen is "starting version 236", followed by "arch-root.....", after which I notice the flashing webcam led again...
Looks like part of the software is always running, and part not. Or is there a problem with the screen, which sometimes is not addressed correctly? The 'starting' message is telling me that the kernel is being loaded, right? So if I don't get that message, why would the webcam still be initialized??
I always update weekly, what would the lts kernel number be? I am running as of today 4.14.13-1-ARCH.
How would I use debug intramfs? I'll try to ping my pc when it is stuck next time, but it might not work since I am not sure of the (dynamic) ip-address.
Finally, my /proc/cmdline:
BOOT_IMAGE=/boot/vmlinuz-linux root=UUID=fc5e2ec5-5d9b-492e-96dc-2ea10141983c rw quiet
Checked the UUID, which is correct.
Offline
https://www.archlinux.org/packages/core … linux-lts/
https://www.archlinux.org/packages/extr … 340xx-lts/
https://wiki.archlinux.org/index.php/Ge … t_problems
If you have another machine, you should check whether you can ping the "failed" one (and ideally setup an ssh daemon there so you can ssh into and live-inspect it)
This is very most likely the GPU so you may also want to try whether this is reproducible w/ nouveau (although I would have said that this should not happen w/ nomodeset then ....)
Offline
Ok, thanks, will try to ping my machine at next failure, and read your suggestions (links).
I also get the feeling that the Nvidia chip is giving problems - it seems to be running quite hot: normal temp is around 80 degrees C, but moving a window around (ok, not like you would do normally, but same effect gives watching a youtube video or playing a game with some video action) quickly rises it to 100 degrees C.
My laptop is because of this permanently on a cooling pad with 2 fans - suspect it to be unlikely though that at boot the temp is high already.
I had worse temp problems a few years ago running another distro which used the nouveau driver, so I forgot about using that driver. How do I switch drivers in Arch?
Last edited by luigi01 (2018-01-22 20:21:49)
Offline
Uninstalling the nvidia packages and maybe installing xf86-video-nouveau should do (w/o the latter X11 will use the modesetting driver, but the kernel will run on nouveau)
GPUs can take a lot of heat, but the suggested values are too high (and also imply massive power drain) - you should check dmesg for errors.
Offline
That's strange - I don't seem to be able to delete the nvidia package:
sudo pacman -R nvidia returns "error: package not found", same when I try to remove nvidia-utils ???
sudo pacman -Qs nvidia gives the following:
local/libvdpau 1.1.1+3+ga21bf7a-1
Nvidia VDPAU library
local/nvidia-340xx 340.106-1
NVIDIA drivers for linux, 340xx legacy branch
local/nvidia-340xx-utils 340.106-1
NVIDIA drivers utilities
local/xf86-video-nouveau 1.0.15-2 (xorg-drivers)
Open Source 2D acceleration driver for nVidia cards
(so there is the xf86 driver I already installed primer to deleting the nvidia driver).
Even typing the full nvidia-340.106-1 does not do the trick - can you tell me what I am doing wrong?
Offline
How did you install archlinux, that package in particular?
Please paste your entire pacman log.
Offline
I installed Arch using the installation guide; installed nvidia using the wiki page. I also do get frequently updates for nvidia (340xx).
Was attracted to Arch for its simplicity, running the very latest software, and being in control rather than pressing a button (like in windows) to run some software to solve your problem without having a clue what is going on.
Sorry to ask, but how do I upload my pacman log file, and show a link here, it is rather large (1.1M) to just copy and paste it here.
I just read the nouveau wiki page, describing how to unload nvidia, and load nouveau, so I could go ahead and give that a try (while watching the gpu temp...).
So far starting has been ok this week.
Last edited by luigi01 (2018-01-25 20:18:48)
Offline
https://pastebin.com
You can also just https://wiki.archlinux.org/index.php/Li … in_clients
Offline
Ok, here we go after some setbacks (file size too large...):
had to split up in 3 files
1: https://pastebin.com/5EAbn6Xn
2: https://pastebin.com/8dkxmjUk
3: https://pastebin.com/RHKi68c8
Offline
Those pastes are private.
Offline
And please use a different pastebin client.
Note: pastebin.com is blocked for some people and has a history of annoying issues (javascript, adverts, poor formatting, etc). Do not use it.
Last edited by 2ManyDogs (2018-01-25 21:17:33)
Offline
After preparing all sorts of additions to the kernel boot command, so I am ready to try to get more data after rebooting when the system fails again, my system behaved like a charm! (like the toothache that is no longer there once you get to the dentist...). I hope with getting all the updates meanwhile the problem was solved.... When I face the same error again I'll try to get more relevant data, and will post it. Thank you all for thinking with me on this.
Offline
Pages: 1