Bug 18603 - Booting fails after boot.iso install
Summary: Booting fails after boot.iso install
Status: RESOLVED OLD
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: Cauldron
Hardware: All Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Mageia Bug Squad
QA Contact:
URL:
Whiteboard:
Keywords: NEEDINFO
Depends on:
Blocks:
 
Reported: 2016-06-01 19:47 CEST by William Kenney
Modified: 2016-07-03 18:57 CEST (History)
6 users (show)

See Also:
Source RPM: nvidia-current-361.45.11-1.mga6.nonfree?
CVE:
Status comment:


Attachments
M6 x86_64 boot.iso install stopping point (441.15 KB, image/jpeg)
2016-06-01 19:48 CEST, William Kenney
Details
bug command result (457.14 KB, image/jpeg)
2016-06-01 21:08 CEST, William Kenney
Details
dmesg command result (398.74 KB, image/jpeg)
2016-06-01 21:09 CEST, William Kenney
Details
m6_report.bug.xz.zip (260.36 KB, application/zip)
2016-06-02 18:20 CEST, William Kenney
Details
m6_Xorg.0.log (26.41 KB, text/x-log)
2016-06-02 18:21 CEST, William Kenney
Details
There's at least a couple failed boots in this journal (608.31 KB, application/x-xz)
2016-06-03 04:42 CEST, William Kenney
Details
logs from boot.iso today (487.77 KB, application/zip)
2016-06-05 19:42 CEST, William Kenney
Details

Description William Kenney 2016-06-01 19:47:22 CEST
Description of problem:

Boot.iso  05/20/16  x86_64  Plasma

After a very long install process using boot.iso
the install basically stopped. It seemed to get
through the nvidia install.
Repo source mirrors.kernel.org

See attached image
Comment 1 William Kenney 2016-06-01 19:48:56 CEST
Created attachment 7884 [details]
M6 x86_64 boot.iso install stopping point
Comment 2 Marja Van Waes 2016-06-01 20:13:55 CEST
Hi William,

Thx for testing again.

As always for classical installs (= non-Live installs), please fetch the logs
https://wiki.mageia.org/en/Triage_guide#Traditional_installer

I understand you were past the partitioning step, so please put report.bug on a USB key as described in the link above.

The file is probably very large. Please compress it with

    xz report.bug

and then attach the resulting report.bug.xz to this report.

Cheers,
Marja

Keywords: (none) => NEEDINFO
CC: (none) => ennael1, marja11, thierry.vignaud, tmb

Comment 3 William Kenney 2016-06-01 21:08:37 CEST
Created attachment 7885 [details]
bug command result
Comment 4 William Kenney 2016-06-01 21:09:02 CEST
Created attachment 7886 [details]
dmesg command result
Comment 5 Thierry Vignaud 2016-06-01 21:34:41 CEST
This is a commmand to run while in installer not after rebooting the installed OS...
Thierry Vignaud 2016-06-01 21:34:46 CEST

Attachment 7885 is obsolete: 0 => 1

Thierry Vignaud 2016-06-01 21:35:07 CEST

Attachment 7886 is obsolete: 0 => 1

Comment 6 Charles Edwards 2016-06-01 22:10:57 CEST
Is this bug about the actual installation process Or about the reboot After the installation?

I did several BIOS installs yesterday using nonfree iso and all completed normally.

An hour ago, with the same iso, I did an EFI install and it, also, completed normally. 


If instead this is about all the audit warnings being displayed.
They are caused by /proc/sys/kernel/printk being set as
7 4 1 7
change it to
4 4 1 7

To make the change I use
# echo 4 4 1 7 > /proc/sys/kernel/printk

CC: (none) => cae

Comment 7 William Kenney 2016-06-02 00:45:24 CEST
(In reply to Charles Edwards from comment #6)

> Is this bug about the actual installation process Or about the reboot After
> the installation?

Opppssss sooooo sorry. Yes, the 1st and 2nd stages of the boot.iso install
went just fine. Fed in the root password, user name and password and set
up the nvidia driver ( new set up ) and all that went just fine. This
issue occurs after the reboot of all of that.
Comment 8 Marja Van Waes 2016-06-02 08:21:55 CEST
(In reply to William Kenney from comment #7)
> (In reply to Charles Edwards from comment #6)
> 
> > Is this bug about the actual installation process Or about the reboot After
> > the installation?
> 
> Opppssss sooooo sorry. Yes, the 1st and 2nd stages of the boot.iso install
> went just fine. Fed in the root password, user name and password and set
> up the nvidia driver ( new set up ) and all that went just fine. This
> issue occurs after the reboot of all of that.

Thanks Charles, and no problem, William.

William, if you can switch to TTY2, please try to get

    /var/log/Xorg.0.log


If switching to a TTY is impossible, can you please retry, but append a 3 to the kernel options in your bootloader choice screen. That should lead you to a login prompt. 

(Btw, when you can do a txt login, _after_ having fetched that log file, you can also try to run XFdrake as root and select another video driver)

Please do also attach 

   /root/drakx/report.bug.xz

Summary: Today's boot.iso install stopping point => Booting fails after boot.iso install

Comment 9 Anne Nicolas 2016-06-02 14:20:47 CEST
Do you use EFI ?if not then grub legacy is automatically installed and you may have been killed by https://bugs.mageia.org/show_bug.cgi?id=18583
Comment 10 Frank Griffin 2016-06-02 15:18:47 CEST
Are you sure this isn't https://bugs.mageia.org/show_bug.cgi?id=17584 ?

CC: (none) => ftg

Comment 11 William Kenney 2016-06-02 17:34:06 CEST
(In reply to Anne Nicolas from comment #9)

> Do you use EFI ?if not then grub legacy is automatically installed and you
> may have been killed by https://bugs.mageia.org/show_bug.cgi?id=18583

I did not use EFI. I'll see if I can get the logs out of it today.
Comment 12 William Kenney 2016-06-02 17:35:02 CEST
(In reply to Frank Griffin from comment #10)

> Are you sure this isn't https://bugs.mageia.org/show_bug.cgi?id=17584 ?

I didn't see the "looking for new hardware" indication.
Comment 13 Thierry Vignaud 2016-06-02 17:51:43 CEST
We would need the output of journalctl
eg: "journalctl --since 2016-05-31 | xz >/tmp/journal.txt.xz"
Comment 14 William Kenney 2016-06-02 18:20:57 CEST
Created attachment 7891 [details]
m6_report.bug.xz.zip
Comment 15 William Kenney 2016-06-02 18:21:41 CEST
Created attachment 7892 [details]
m6_Xorg.0.log
Comment 16 William Kenney 2016-06-02 18:27:23 CEST
(In reply to Marja van Waes from comment #8)

> William, if you can switch to TTY2, please try to get
> 
>    /var/log/Xorg.0.log.....
>    /root/drakx/report.bug.xz

Just for yuks here's how I got these two files. I loaded up the
failing drive but booted the system using the M5 x86_64 KDE Live-DVD.
Then using an "su -" then kwrite/dolphin commands in a terminal I was
able to copy/re-create the files in the /home/live/Documents directory
of the Live-DVD. I then copied them on to a USB drive. :-))
Comment 17 Charles Edwards 2016-06-02 22:53:45 CEST
If the nivida driver is selected during the installation the modules is not built
until 1st boot After the installation is completed, and this can make it seem that the boot has stopped|failed which Is Not the case.

Did another test install today.
CPU: Phenom II x4 970, 3.49GHz
GPU: Geforce 560GTX

The kernel used in this installation kernel-desktop-4.6.1-1 and there were no issues during the installation.

I use splash=verbose and boot to init 3 

On first boot I see the notice that the nvidia module build has started.
After a few seconds I see 2 additional prompt entries.
The 2nd entry is "24 audit callbacks suppressed"
(IIRC they were not suppressed in previous kernels)

Next at the prompt I see 
A timer that the dkms build is in progress with no end limit set.
On this system it took 2 min for the module build to complete.
The boot then continued to the login prompt.

Even at the login prompt audit call back are still be given until I "fix" printk
as described in Comment 6.
Comment 18 William Kenney 2016-06-03 04:42:01 CEST
Created attachment 7894 [details]
There's at least a couple failed boots in this journal
Comment 19 William Kenney 2016-06-03 04:43:31 CEST
(In reply to William Kenney from comment #18)

> Created attachment 7894 [details]
> There's at least a couple failed boots in this journal

Added journal.txt.xz
Comment 20 William Kenney 2016-06-05 19:42:57 CEST
Created attachment 7918 [details]
logs from boot.iso today

logs on 5 June 16

included are journalctl.txt ( one unsuccessful boot )
all the files in /root/drakx/*.*

there was no /var/log/Xorg.0.log
Comment 21 Thierry Vignaud 2016-06-17 11:30:41 CEST
That's not installer bug.

You did a minimal install with x11?
With proprietary nvidia driver.

Jun 05 10:00:32 localhost systemd[1]: prefdm.service: Main process exited, code=exited, status=1/FAILURE
Jun 05 10:00:32 localhost systemd[1]: prefdm.service: Unit entered failed state.
Jun 05 10:00:32 localhost systemd[1]: prefdm.service: Triggering OnFailure= dependencies.
Jun 05 10:00:32 localhost systemd[1]: prefdm.service: Failed with result 'exit-code'.

Maybe the dkms nvidia driver failed to build.
Could you get /var/log/Xorg.0.log or ~/.local/share/xorg/Xorg.0.log

Component: Installer => RPM Packages
Source RPM: (none) => nvidia-current-361.45.11-1.mga6.nonfree?

Comment 22 Thomas Backlund 2016-06-17 11:36:11 CEST
without checking the logs...

is there a "nokmsboot" on the kernel command line ?

If not, thats the reason nvidia fails
Comment 23 Thierry Vignaud 2016-06-17 11:47:09 CEST
There's not (only "root=/dev/sda1 ro").

@William: As this was done with drakx 17.33 whereas there were grub2 fixes after, can you retry with current cauldron?
Comment 24 Marja Van Waes 2016-07-03 06:39:01 CEST
(In reply to Thierry Vignaud from comment #23)
> There's not (only "root=/dev/sda1 ro").
> 
> @William: As this was done with drakx 17.33 whereas there were grub2 fixes
> after, can you retry with current cauldron?

Is it still valid with 6sta1, wilcal?
Comment 25 Marja Van Waes 2016-07-03 06:39:36 CEST
s/6sta1/current cauldron/
Comment 26 William Kenney 2016-07-03 06:49:25 CEST
It's been awhile since I've done a boot.iso install. I'll execute one
tomorrow ( 3 July ) California time.
Comment 27 William Kenney 2016-07-03 18:57:16 CEST
I'm going to set this bug to "resolved/old" in lieu of:

https://bugs.mageia.org/show_bug.cgi?id=18852

Status: NEW => RESOLVED
Resolution: (none) => OLD


Note You need to log in before you can comment on or make changes to this bug.