On 06/13/18 10:47, Alex wrote:
Hi,
On Tue, Jun 12, 2018 at 6:16 PM, Ed Greshko <ed.greshko(a)greshko.com> wrote:
> On 06/13/18 04:45, Alex wrote:
>> Hi,
>> I have a fedora27 system on an older x86_64 machine and when booting
>> the latest kernel (I believe it was just released today) as well as
>> the previous two, the display goes dark and becomes unresponsive.
>>
>> I've booted into the rescue kernel (an older 3.14 from fc21) to get it
>> back online.
>> $ uname -a
>> Linux localhost 3.17.4-301.fc21.x86_64 #1 SMP Thu Nov 27 19:09:10 UTC
>> 2014 x86_64 x86_64 x86_64 GNU/Linux
>>
>> All other current kernels failed.
>>
>> How do I troubleshoot this?
>>
>> Based on the output from journalctl, it appears to be something
>> related with gnome and/or packagekit and/or the display driver. Here's
>> a few select error messages that look like they could be the cause.
>> I've also uploaded the journalctl boot entry that I believe is
>> associated with this boot to pastebin:
>>
>>
https://pastebin.com/804NfcEB
>>
> This does not appear to be a journal from a failed boot.
My apologies; can we try again?
https://pastebin.com/BjKf0Kdt
That can't be a "crash" either because....
Jun 12 15:52:27 localhost systemd[1]: Reached target Graphical Interface.
The Graphic Target has been reached...
Jun 12 15:52:28 localhost systemd[1]: Startup finished in 2.883s (kernel) + 4.719s
(initrd) + 1min 44.210s (userspace) = 1min 51.812s.
The startup is complete.
Jun 12 15:52:29 localhost dbus-daemon[1431]: [session uid=1000 pid=1426] Successfully
activated service 'org.gnome.evolution.dataserver.Calendar7'
Jun 12 15:52:29 localhost dbus-daemon[1430]: [session uid=1000 pid=1427] Successfully
activated service 'org.gnome.evolution.dataserver.Calendar7'
UID 1000 seems to be set for automatic login and evolution processes are starting.
Jun 12 15:52:35 localhost sshd[2954]: Invalid user tech from 42.99.180.167 port 57531
And sshd is running and someone has tried to login....
Is it possible the system crashes prior to logging? I'm assuming
the
current boot is 0, which is the rescue kernel. The previous boot was
the crash with the latest kernel, so I'm assuming that's -1, which is
what's posted in the pastebin above.
Yes, it is possible you won't get to logging....
If you look at the very start of a boot you should see something like this...
[egreshko@acer ~]$ journalctl -b | more
- Logs begin at Fri 2017-11-24 06:58:33 CST, end at Wed 2018-06-13 11:05:48 CST. -
-
Jun 13 10:38:19
acer.greshko.com kernel: microcode: microcode updated early to revi
sion 0xa4, date = 2010-10-02
Jun 13 10:38:19
acer.greshko.com kernel: Linux version 4.16.14-300.fc28.x86_64 (moc
kbuild(a)bkernel02.phx2.fedoraproject.org) (gcc version 8.1.1 20180502 (Red Hat 8.1.1
-1) (GCC)) #1 SMP Tue Jun 5 16:23:44 UTC 2018
Jun 13 10:38:19
acer.greshko.com kernel: Command line: BOOT_IMAGE=/vmlinuz-4.16.14-
300.fc28.x86_64 root=/dev/mapper/fedora_acer-root ro rd.lvm.lv=fedora_acer/root rd.
lvm.lv=fedora_acer/swap rhgb quiet nouveau.modeset=0 rd.driver.blacklist=nouveau vi
deo=vesa:off LANG=en_US.UTF-8
Jun 13 10:38:19
acer.greshko.com kernel: x86/fpu: x87 FPU will use FXSAVE
Jun 13 10:38:19
acer.greshko.com kernel: e820: BIOS-provided physical RAM map:
Jun 13 10:38:19
acer.greshko.com kernel: BIOS-e820: [mem 0x0000000000000000-0x00000
0000009f7ff] usable
> Your subject seems to imply that you never get to login screen at either a graphical
> or run level 3 boot.
Yes, that's correct. There are a lot of "failure" and other
"fail"
errors including "Failed to get current display".
> Suggest you also edit your boot parameters to remove "rhgb" and
"quiet" to watch the
> boot process more closely.
I've done that now. This is actually a friend's computer, located
about 30 miles away, so I can't reboot it very easily without making
sure he's there to monitor.
OK. Someone has to be at the system to see exactly what is going on....
Samuel Sieb wrote:
> Can you access it remotely using ssh?
No, it's completely catatonic - no ping, no ssh, no disk activity.
And, as you can see from the above, the info you've put on pastebin shows sshd to be
running and accepting connections.
--
Conjecture is just a conclusion based on incomplete information. It isn't a fact.