PCLinuxOS Doesn't always boot successfully - SOLVED

[mag]

About 1 time in 5 PCLOS 2011 doesn't boot successfully.

I get the grub screen ok, and then the booting progress bar, but instead of going to the start-up (splash?) screen it then just goes blank. and the monitor displays 'no signal'.

I have to shutdown using the power button.

Any thoughts on what might be wrong/how to fix (that a relative newbie might be able to follow).

Thanks

mag

[Hootiegibbon]

This sounds like you need to check your monitor settings in "configure your computer" >hardawre >graphics card as the sync details are set incorrectly

Jase

[mag]

Thanks.

This is an old system (e-machines/emprex - not apparently a listed vendor for monitors), and I don't have the original details - but re-running the monitor set-up has sorted another problem I had with the windows being ever-so-slightly too big for the screen.

The intermittent fail to boot problem hasn't recurred since either (fingers crossed).

[mag]

Hello again,

The problem is still there (and still intermittent).

Could you clarify what you meant by ' as the sync details are set incorrectly' please?

The monitor says Emprex LM-1704 on the back.

Under 'configure video card' settings the monitor has been detected as LM-1704 (plug'n play). So I'm not sure what there is to sync

The video card is shown as ATI Radeon x1950 and earlier.

[pmpatrick]

If you have an emachines box, your problem may be traceable to the rotten, cheap Bestec power supplies they frequently put in those things. These PSUs can cause all kinds of erratic problems and, as an added bonus, when they blow, they usually take out more than a few components with them. Boot time is a time of high stress for a PSU as all components are powering up. With a marginal or weak PSU, you can see boot time problems like you are experiencing.

If you have a spare power supply around, I would recommend swapping it in for the Bestec or other branded cheap PSU that emachines habitually uses. Erratic boot time behavior may well indicate the PSU is going, and, as noted above, those cheap emachine PSUs do not fail gracefully.

[longtom]

Just throwing a curveball here:

How full is your / partition?

[CaptainSarcastic]

If you have an emachines box, your problem may be traceable to the rotten, cheap Bestec power supplies they frequently put in those things. These PSUs can cause all kinds of erratic problems and, as an added bonus, when they blow, they usually take out more than a few components with them. Boot time is a time of high stress for a PSU as all components are powering up. With a marginal or weak PSU, you can see boot time problems like you are experiencing.

If you have a spare power supply around, I would recommend swapping it in for the Bestec or other branded cheap PSU that emachines habitually uses. Erratic boot time behavior may well indicate the PSU is going, and, as noted above, those cheap emachine PSUs do not fail gracefully.

I can second this.  I've had good success with the socket 939 eMachines, but the first thing I've done with them is yank the PSU and replace it.  The Intel-based eMachines seemed more failure-prone than the AMD-based ones in general, too.

To the OP, is this is a fresh install?  Did you install the KDE edition of PCLOS?  Was this machine reliably running another OS prior to installing PCLOS?  Does booting from a Live CD have the same success/failure rate?

[mag]

Thanks for the responses.

Trying to provide the info requested:

How full is your / partition?

38%

is this is a fresh install?  

Yes (freshish - maybe  4 weeks old)

Did you install the KDE edition of PCLOS?  

Yes

Was this machine reliably running another OS prior to installing PCLOS?  

This is an old XP machine that became slow (but always booted OK - slowly - and ran reliably). I got a new W7 machine, and decided to dedicate this one to playing with Linux (about which I know very little - but I liked the little I'd heard)

First I installed Mint 8, which ran very successfully - but was not supported. I then tried Mint 11 and LMDE and another distro 10 and 11- none would boot without ACPI tweaks that seemed to cause fan speed issues. With the ACPI tweaks they did boot OK though.

PCLinuxOS was the first major distro I tried that installed, booted and ran without fuss.

Does booting from a Live CD have the same success/failure rate?

I'm afraid I couldn't really say, because I don't boot from PCLinuxOS from LiveCD often enough to be statistically relevant. I have booted Slitaz 3 (a tiny distro that runs entirely in memory) from LiveCD quite a few times and never had an issue

Hope this helps.

(Sometimes when I come to a Linux forum it feels like I'm in France - I speak the language just well enough to ask questions, but not well enough to follow the answers)

Edit: I tried a bit of investigation. I got rid of quiet splash from grub to see where it was happening. It always happens after the message:

acpid waiting for event: event logging is off (or something like that)

(how would I turn this event logging on by the way?)

I also tried Alt/SysRq REISUB to try to safely shut down - no joy.

[melodie]

Hi Mag,

I don't know what goes on with your machine, but maybe it has not enough memory to run KDE or any heavy desktop ? If you don't know Linux : Linux has many desktops - we use mostly KDE and Gnome, then more and more often Xfce, Lxde, and since a pair of years we also use Openbox -almost standalone-

KDE and Gnome are the ones which require the more resources, Xfce less, and almost as generous in features as KDE, the Lxde is next, less greedy, and becoming more and more easy to use, with many graphical use interfaces to configure your preferences. Openbox standalone is to be left for people already knowing Linux a little.

If you want to check the memory, you could open a console and type "free" hit return, then "free -m" hit return. Please bring back the output.

If you want to test the memory, you could run a memtest (an option which is available at boot).

I would also suggest you look at what graphical driver is set : I suppose it would be radeon ? You can do this in console:

grep "Driver" /etc/X11/xorg.conf

and if you open the PCC at the Hardware>Video section, look if hardware acceleration is on : if so, set it to off (in the advanced part of that section). If hardware acceleration is on for the mouse, set it to off as well.

I also tried Alt/SysRq REISUB to try to safely shut down - no joy.

Does that mean it didn't work ? If so, you can enable it in the file /etc/sysctl.conf : the line "kernel.sysrq = " should be set to 1, this way : kernel.sysrq = 1

Regards,
Mélodie

[Hootiegibbon]

Hello again,

The problem is still there (and still intermittent).

Could you clarify what you meant by ' as the sync details are set incorrectly' please?

The monitor says Emprex LM-1704 on the back.

Under 'configure video card' settings the monitor has been detected as LM-1704 (plug'n play). So I'm not sure what there is to sync

The video card is shown as ATI Radeon x1950 and earlier.

This sounds like a case that EDD is not receiving the correct details from your monitor.

This would also account for no visual evidence that the magic syskeys were working

Set the monitor via configure your computer not to use plug and play but choose either vendor (and pick your monitor if there) or use the generic setting for the monitor type/natice resolution

then edit (as root) the menu.lst file to add "edd=off" to the end of the default bootline this will mean you have manually set up teh monitor correctly and instructed the system not to probe and reset teh monitor on boot.

This will hopefully solve the issue

Jase

[mag]

Again, many thanks for the help and advice.

@Hootiegibbon
I have done what you suggested (I hope) - will report back after a few more reboots.

@Melodie
The machine has 1.5 gig of RAM - so should be OK I think.
             total       used       free     shared    buffers     cached
Mem:       1421476     795260     626216          0      25836     380432
-/+ buffers/cache:     388992    1032484
Swap:      4088504          0    4088504
[dg@localhost ~]$ free -m
             total       used       free     shared    buffers     cached
Mem:          1388        776        611          0         25        371
-/+ buffers/cache:        379       1008
Swap:         3992          0       3992
The driver is just shown as 'ati'

I'll try out the other stuff and get back to you.

[mag]

OK, I set kernel.sysrq = 1 (hope I don't need to use it any time soon). Hardware acceleration on video was off.

I have rebooted six or seven times without incident, so I'm hoping this issue is now fixed (but being a cautious soul by nature I'll wait a few days before marking the thread SOLVED).

Many thanks again to all.

mag

[mag]

Unfortunately the occasional fail to boot problem persists, despite setting edd=off.

I have also tried switching to PCLOS Openbox - but that hasn't affected it (not surprising I guess, since the problem occurs before login screen).

The failure always seems to occur just after the 'Starting display manager' message.

Any further thoughts appreciated.

[melodie]

Hi,

Let's try to get more info from your hardware. Please do as root in console:

Code: [Select]

lspci -v | grep VGA
and bring back the result.

Tell us if the video card is embedded. If yes, could you replace it with another one ? Any other brand, external pci would be ok. Perhaps the card is sending wrong info to the screen.

[mag]

Thanks for the reply.

This brought back:

lspci -v | grep VGA
01:05.0 VGA compatible controller: ATI Technologies Inc RC410 [Radeon Xpress 200] (prog-if 00 [VGA controller])

I'm afraid I don't know how I would tell if the video card is embedded (or what this means). I don't have any other video cards.

(Edit - after a quick wiki I think it probably is an embedded card - ie it borrows some of the main system's RAM).