[SOLVED] After kernel-update ... only restarts

[aldQueiroz]

HI everybody.
By suggestion of a friend, very experienced with Linux, I've upgraded my desktop running Mageia 3 from kernel-desktop 3.8.13-4 to the brand new 3:10:19 , available in media ' Testing ' by now; the system got actually faster , with less latency boot ! But ... I've got an annoying problem, again: the desktop does not turn off , it only restarts , no matter if trying to shut down by the 'button ' on the panel ', by the options the launcher menu, or even via terminal ( command ' shutdown- h now ' or ' poweroff ' , logged as root ) ; shutdown failures to interrupt some service I could not identify ....
I've had this problem before , when I was using Ubuntu LTS 10:04 - once a time, there was a routine kernel update ( to kernel 2.6.32-31 ); the solution was to boot with an older kernel ( 2.6.32-30 ) , but each time there was a newer update , I had to redo the grub configuration to start with the older kernel ... I never realized how to get rid of that. Further, when Ubuntu introduced 'Unity' , which i disliked very much, that's why I changed to Mageia ...
Indeed, I had this very problem weeks ago, after a kernel update to kernel-desktop-3.8.13-4 (I think...) from core repositories, and the ' by-pass ' was the same as used before with Ubuntu : Configure default boot to use the older kernel 3.8.13-1... But this is not a proper solution !

My desktop configurations:
CPU Athlon 64 x2 , motherboard PHITRONICS AM2N1K-M V3.0, Bios Phoenix 4.071.032/02/07;
video card 'onboard' C61 NVIDIA (GeForce 6100 nForce 405], nvidia driver (did not catch the version ...);
soundcard 'onboard' NVIDIA MCP61 High Definition Audio, snd_hda_intel driver ;
The 'systemd', 'acpi' and 'acpid' packages are installed (v. v. 195.22-1.m3, 1.6 and 2.0.18, respectively) and running (MCC >> Advanced >> Startup ... check!);
Bios setup shows that ACPI function is active, but it doesn't give me option to turn on / off ...;

In another forum people asked me to post messages from var/log/messages (through terminal : ' # less /var/log/messages ') and from other files as well:
messages.log
kernel.log
syslog.log
I tried. But, honestly, what I've got seems like 'crap' to me (a very, very long output, and without anything I could identify to be related with the no shutdown problem... I couldn't either save it as txt file!)
Other suggestion was to investigate MSEC setup -- one of its possible configurations should impeach shutdown, according to one post... But, again, I have no idea of which parameter would related to this (I checked MSEC in my desktop, and it's set to default security levels...)

Anyone have another idea ?

Waiting for some reply and help.
Thank you !

[daniewicz]

During the boot process you should have a means of choosing which kernel you would like to use. Can you not choose the older kernel at this point?

[aldQueiroz]

Yes,I can choose a former kernel. But that's not the point.
The new kernel 3.10.19 performs very well, indeed -- everything got faster, including boot process. The only problem so far is this issue when turning off; so, I don't intend to return to an older kernel...
Further, actually I've experienced the same problem with a recent update of the kernel available from 'core' repository: 3.8.13-4, I think. I had to revert to kernel 3.8.13-1 to enable proper turn-off again.
I'm away from my machine, during a business travel. Once I return home, I'll try to get some info from the var/log/messages and post it, so one can figure out which process fails to stop.
By the way, I missed to mention that terminal commands " shutdown" and "poweroff" didn't work...
Thanks!

[daniewicz]

OK I understand. We look forward to some var/log/messages.

[aldQueiroz]

Hi guys,
I've been absent from here, without posting for a long while because my desktop was 'unusable'! Precisely, due to an improper shutdown (by another user ...), which led to an error in a file system and stopped booting correctly; the X server did not start right any longer ... (irony: first, my machine did not turn off, and then it didn't turn on! ...). I do not know if it was the video driver, or whatsoever... The most practical solution was to install the whole system again! Fortunately I have installed with "/home" apart of partition "/" ...!
Well, herewith attached there is file system messages log I collected, referred to the problem of 'not shutting down, restarting only' ;

If someone among you be able (and patient enough!) to analyse this, so we could find where is the d... problem!... I've been having this issue since kernel update to version kernel-desktop-3.8.13-1 (from Core updates); 'Testing' kernels 3.10.19 and 3.10.24-2 present the same problem, what it's a pity, 'cause with my machine got faster using such update...

Regards!

[filip]

If someone among you be able (and patient enough!) to analyze this, ...

I'm probably not able to properly analyze it but I did found some strange entries with errors or failures. Extract fileare also attached.

This might be connected:

Code: Select all

Dec  1 23:00:40 Bazinga dbus-daemon[1086]: dbus[1086]: [system] Rejected send message, 2 matched rules; type="method_call", sender=":1.26" (uid=500 pid=2207 comm="kdeinit4: kded4 [kdeinit]                       ") interface="org.freedesktop.login1.Manager" member="CanHybridSleep" error name="(unset)" requested_reply="0" destination="org.freedesktop.login1" (uid=0 pid=1081 comm="/usr/lib/systemd/systemd-logind ")
Dec  1 23:00:40 Bazinga dbus[1086]: [system] Rejected send message, 2 matched rules; type="method_call", sender=":1.26" (uid=500 pid=2207 comm="kdeinit4: kded4 [kdeinit]                       ") interface="org.freedesktop.login1.Manager" member="CanHybridSleep" error name="(unset)" requested_reply="0" destination="org.freedesktop.login1" (uid=0 pid=1081 comm="/usr/lib/systemd/systemd-logind ")

Dec  5 07:24:35 Bazinga dbus[1103]: [system] Rejected send message, 2 matched rules; type="method_call", sender=":1.27" (uid=500 pid=2691 comm="kdeinit4: kded4 [kdeinit]                       ") interface="org.freedesktop.login1.Manager" member="CanHybridSleep" error name="(unset)" requested_reply="0" destination="org.freedesktop.login1" (uid=0 pid=1099 comm="/usr/lib/systemd/systemd-logind ")
Dec  5 07:24:35 Bazinga dbus-daemon[1103]: dbus[1103]: [system] Rejected send message, 2 matched rules; type="method_call", sender=":1.27" (uid=500 pid=2691 comm="kdeinit4: kded4 [kdeinit]                       ") interface="org.freedesktop.login1.Manager" member="CanHybridSleep" error name="(unset)" requested_reply="0" destination="org.freedesktop.login1" (uid=0 pid=1099 comm="/usr/lib/systemd/systemd-logind ")

There are some ata1 entries that gives me unpleasant feeling like this one:

Code: Select all

Dec  1 18:28:48 Bazinga kernel: ata1.00: exception Emask 0x10 SAct 0x0 SErr 0x1950000 action 0xe frozen
Dec  1 18:28:48 Bazinga kernel: ata1: SError: { PHYRdyChg CommWake Dispar LinkSeq TrStaTrns }
Dec  1 18:28:48 Bazinga kernel: ata1.00: failed command: READ DMA
Dec  1 18:28:48 Bazinga kernel: ata1.00: cmd c8/00:08:e8:50:ce/00:00:00:00:00/e0 tag 0 dma 4096 in
Dec  1 18:28:48 Bazinga kernel:          res 40/00:00:00:4f:c2/00:00:00:00:00/00 Emask 0x14 (ATA bus error)
Dec  1 18:28:48 Bazinga kernel: ata1.00: status: { DRDY }
Dec  1 18:28:48 Bazinga kernel: ata1: hard resetting link
Dec  1 18:28:48 Bazinga kernel: ata1: nv: skipping hardreset on occupied port
Dec  1 18:28:49 Bazinga kernel: ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Dec  1 18:28:49 Bazinga kernel: ata1.00: configured for UDMA/133
Dec  1 18:28:49 Bazinga kernel: ata1.00: device reported invalid CHS sector 0
Dec  1 18:28:49 Bazinga kernel: ata1: EH complete

There are also some entries which might tell us that one or more module is not loaded properly:

Code: Select all

Dec  2 13:31:43 Bazinga modprobe[1090]: modprobe: ERROR: missing parameters. See -h.

There seems to be some problems with services also:

Code: Select all

Dec  5 07:20:20 Bazinga systemd[1]: upower.service: main process exited, code=killed, status=5/TRAP
Dec  5 07:20:20 Bazinga systemd[1]: Unit upower.service entered failed state
Dec  5 07:20:20 Bazinga systemd[1]: udisksd.service: main process exited, code=exited, status=1/FAILURE
Dec  5 07:20:20 Bazinga systemd[1]: Unit udisksd.service entered failed state

HTH someone.

system-messages_errors-only.txt

Extract of system-messages.txt

(47.46 KiB) Downloaded 230 times

[aldQueiroz]

Thank you, filip! Now we have some stuff to work with, rather than just keeping try and errors with several kernel versions, and always suitable to see the issue returns in newer kernel updates!
By the way, my friend and 'linux tutor' reported he got this issue too (running Mageia 3 x86-64, in a similar configuration AMD 64 x2) when running kernels 3.10.19 and the latest 3.10.24-2 from Core Updates; kernel 3.10.22 didn't show this issue so far.

Thanks again, and regards to all!

[aldQueiroz]

Hi folks,
Despite have posted the messages log file , unfortunately no objective conclusion was found about why the problem of " not turning-off, just restart ." Again, I experienced the problem using the following kernel -desktop : 3.8.13-4 ( core updates ) , 3.10.19 (testing ), with my 'then configuration: Mageia 3 x86_64 , running on an Athlon 64 x2 with Phoenix Bios (not sure about version of bios : what I read on boot screen before the grub was " 4.071.032/02/07 " , while the command '# dmidecode | more' returns 'PG version 09.03.2007 6:00' ... )

Well , I could not wait a lifetime for a solution ...
So I did two things :
First, I changed the motherboard (and consequently, the bios!) -- the old one Phitronics had a chronic bad contact at RAM slots ... I switched for Gigabyte motherboard, with Award bios ;
Second, I returned to 3.8.13-1 kernel , which does not present the problem in question . But I've already tested the config with the kernel-tmb-desktop-3.10.25 -- from 'testing' repository - which worked well and got no problems ! I may I upgrade my ' production ' system during coming holiday .

So , I'll state this topic as SOLVED , although we have not found exactly what causes the mentioned problem .

Regards.