Mageia forum

I have a mythtv box that has been running well since I installed it under Magei 2 last fall. I've upgraded the kernel a couple of times and the last 2 kernels seem to be less stable -- the box will freeze occasionally. And by freeze I mean it will hang and run the same audio/video frame of about 0.5 seconds over and over until I reboot.

On rebooting, I'm not getting the system back. Initially, it boots to grub and then if I choose the default kernel (vmlinuz-3.4.34-desktop586-1.mga2), or just let the bootloader time out, it goes to a black screen. I can get it to continue booting by choosing the "no apic" kernel option, bu then it hangs up after initializing the legacy usb system.

Choosing a different kernel, for example 3.4.34-desktop-1.mga1, will allow the machine to boot and run ok, but I may or may not have to include the "no acpi" option.

I'm assuming there is probably some kind of an issue between my hardware and this kernel, though initially this system booted and ran fine under that kernel. At any rate, what would you recommend as the best resolution/work around for this? I'm inclined to just use the control center to make kernel 3.4.34-desktop-1.mga2 the default kernel, but I wanted to check here to see if there was some wisdom.

Does anybody have any light to shed on this situation?

Thanks

Mark

A little more info:

It's a 32 bit system with 1Gb of RAM and an 8600 series nvidia Geforce graphics card.
*-core
description: Motherboard
product: Wolfdale1333-D667.
physical id: 0
*-firmware
description: BIOS
vendor: American Megatrends Inc.
physical id: 0
version: P1.30
date: 02/27/2008
size: 64KiB
capacity: 448KiB


*-cpu
description: CPU
product: Intel(R) Pentium(R) 4 CPU 3.40GHz
vendor: Intel Corp.
physical id: 4
bus info: cpu@0
version: 15.3.4
serial: 0000-0F34-0000-0000-0000-0000
slot: CPUSocket
size: 3400MHz
capacity: 3400MHz
width: 32 bits
clock: 200MHz


*-memory
description: System Memory
physical id: 10
slot: System board or motherboard
size: 1GiB

*-display
description: VGA compatible controller
product: G84 [GeForce 8600 GT]
vendor: nVidia Corporation
physical id: 0
bus info: pci@0000:04:00.0
version: a1
width: 64 bits
clock: 33MHz


I'm considering just uninstalling the 586 kernels, but as I recall, the 586 kernels were supposed to function in the widest range of hardware.

Any advice on that idea?

Thanks.

Mark

mark9117 wrote:I'm considering just uninstalling the 586 kernels, but as I recall, the 586 kernels were supposed to function in the widest range of hardware.

Any advice on that idea?

Thanks.

Mark

The 586 kernel is used on the Live CD for really old hardware but if your system boots with the desktop kernel you don't need it.

That still doesn't explain the problem you're having.

Ken-Bergen wrote:

mark9117 wrote:I'm considering just uninstalling the 586 kernels, but as I recall, the 586 kernels were supposed to function in the widest range of hardware.

Any advice on that idea?

Thanks.

Mark

The 586 kernel is used on the Live CD for really old hardware but if your system boots with the desktop kernel you don't need it.

That still doesn't explain the problem you're having.

Agreed. Thanks.

Okay, I removed the desktop586 kernels which left me with 3.4.32-desktop-2.mga2, and 3.4.34-desktop-1.mga2.

Neither of these kernels will boot. They don't get out of grub without "no acpi" and then they do a fair job of booting and hang. Funny thing is, I can ssh into the box and work on it. I can install packages and even access the graphical mcc.

Logs aren't helping me much, though I did find this at the bottom of the /var/log/kernel/info.log:

Code: Select all

Apr  8 00:17:31 pvr kernel: [   39.208776] r8169 0000:01:00.0: eth0: link up
Apr  8 00:17:31 pvr kernel: [   39.208904] ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
Apr  8 00:17:33 pvr kernel: [   40.783540] NET: Registered protocol family 17
Apr  8 00:17:42 pvr kernel: [   50.114046] eth0: no IPv6 routers present
Apr  8 00:17:42 pvr kernel: [   50.509875] parport0: PC-style at 0x378 (0x778) [PCSPP,TRISTATE]
Apr  8 00:17:42 pvr kernel: [   50.510489] parport0: irq 7 detected
Apr  8 00:17:42 pvr kernel: [   50.592495] ppdev: user-space parallel port driver
Apr  8 00:22:42 pvr kernel: [  349.995002] CPU0: Core temperature/speed normal
Apr  8 00:26:52 pvr kernel: [  600.000031] [Hardware Error]: Machine check events logged


No idea what "Hardware Error" it's seeing.

Any idea on how I can even begin to fix this without a reinstall?

Thanks.

Mark

Try editing your menu entry so you point directly rather than the sym-links

example from

Code: Select all

title mageia linux
kernel (hd0,x)/boot/vmlinuz BOOT_IMAGE=linux root=UUID=xxxxxxxxxxxxxxxxxxxx  resume=UUID=xxxxxxxxxxxxxxx nokmsboot splash quiet vga=788
root (hd0,x)
initrd /boot/initrd.img

to

Code: Select all

title mageia linux
kernel (hd0,x)/boot/vmlinuz-3.4.34-desktop-1.mga2 BOOT_IMAGE=linux root=UUID=xxxxxxxxxxxxxxxxxxxx  nokmsboot noapic vga=788
initrd (hd0,x)/boot/initrd-3.4.34-desktop-1.mga2.img


Don't forget nokmsboot and noapic (for a start, then remove it (noapic) to check if it still boots). Add back the swap later on (resume=UUID=xxxxxxxxxx).
Then add back quiet, splash (if you like) and if this crash it........uh, let us know.

Okay, funny things afoot.

I came home today and found the htpc not running. I turned it on and let it boot without any intervention. It started okay and ran ok.

At that point, I checked out this post and started twiddling. Rather than edit the existing entries in the menu.lst, I added the last stanza you see below:

Code: Select all

timeout 10
color black/cyan yellow/cyan
gfxmenu (hd0,0)/boot/gfxmenu
default 0

title linux
kernel (hd0,0)/boot/vmlinuz BOOT_IMAGE=linux root=UUID=1cc2c773-81a9-4f30-adad-eb1096c0129f splash quiet resume=UUID=404e8708-9a22-4f6f-a614-3317a031b72e PROFILE=home vga=788
initrd (hd0,0)/boot/initrd.img

title linux-nonfb
kernel (hd0,0)/boot/vmlinuz BOOT_IMAGE=linux-nonfb root=UUID=1cc2c773-81a9-4f30-adad-eb1096c0129f resume=UUID=404e8708-9a22-4f6f-a614-3317a031b72e
initrd (hd0,0)/boot/initrd.img

title failsafe
kernel (hd0,0)/boot/vmlinuz BOOT_IMAGE=failsafe root=UUID=1cc2c773-81a9-4f30-adad-eb1096c0129f nokmsboot failsafe
initrd (hd0,0)/boot/initrd.img

title desktop 3.4.32-2.mga2
kernel (hd0,0)/boot/vmlinuz-3.4.32-desktop-2.mga2 BOOT_IMAGE=desktop_3.4.32-2.mga2 root=UUID=1cc2c773-81a9-4f30-adad-eb1096c0129f splash quiet resume=UUID=404e8708-9a22-4f6f-a614-3317a031b72e vga=788
initrd (hd0,0)/boot/initrd-3.4.32-desktop-2.mga2.img

title desktop 3.4.34-1.mga2
kernel (hd0,0)/boot/vmlinuz-3.4.34-desktop-1.mga2 BOOT_IMAGE=desktop_3.4.34-1.mga2 root=UUID=1cc2c773-81a9-4f30-adad-eb1096c0129f splash quiet resume=UUID=404e8708-9a22-4f6f-a614-3317a031b72e vga=788
initrd (hd0,0)/boot/initrd-3.4.34-desktop-1.mga2.img

title Test desktop 3.4.34-1.mga2
kernel (hd0,0)/boot/vmlinuz-3.4.34-desktop-1.mga2 BOOT_IMAGE=desktop_3.4.34-1.mga2 root=UUID=1cc2c773-81a9-4f30-adad-eb1096c0129f nokmsboot noapic resume=UUID=404e8708-9a22-4f6f-a614-3317a031b72e vga=788
initrd (hd0,0)/boot/initrd-3.4.34-desktop-1.mga2.img


I tried to boot that Test line and it failed - went from there to a black screen, no light activity on the computer case, it just stopped.

I rebooted to the default line ("linux") and it booted okay, appeared to hang up at one point (don't remember where) but eventually got passed it and loaded up as expected. I've rebooted it twice since then with no trouble, and it was not necessary to set the "noacpi" kernel option.

The only thing I'm finding in the logs and messages at boot seem to indicate that things are running a little hot. Sensors are showing me this:

Code: Select all

w83627ehf-isa-0290
Adapter: ISA adapter
Vcore:        +1.26 V  (min =  +0.00 V, max =  +1.74 V)
in1:          +1.29 V  (min =  +2.04 V, max =  +1.96 V)  ALARM
AVCC:         +3.30 V  (min =  +2.98 V, max =  +3.63 V)
+3.3V:        +3.28 V  (min =  +2.98 V, max =  +3.63 V)
in4:          +1.67 V  (min =  +2.04 V, max =  +2.03 V)  ALARM
in5:          +1.70 V  (min =  +2.04 V, max =  +1.75 V)  ALARM
in6:          +1.86 V  (min =  +2.04 V, max =  +1.51 V)  ALARM
3VSB:         +3.28 V  (min =  +2.98 V, max =  +3.63 V)
Vbat:         +3.18 V  (min =  +2.70 V, max =  +3.63 V)
in9:          +1.62 V  (min =  +1.94 V, max =  +2.04 V)  ALARM
fan1:        2636 RPM  (min =  664 RPM, div = 8)
fan2:        2220 RPM  (min =  666 RPM, div = 8)
fan3:           0 RPM  (min =    0 RPM, div = 128)
fan5:           0 RPM  (min =    0 RPM, div = 128)
temp1:        +53.0°C  (high = -17.0°C, hyst = -65.0°C)  ALARM  sensor = thermistor
temp2:        +68.0°C  (high = +80.0°C, hyst = +75.0°C)  sensor = thermistor
temp3:        +50.0°C  (high = +80.0°C, hyst = +75.0°C)  sensor = thermistor
cpu0_vid:    +0.000 V


I'm no hardware engineer, but I suspect a lot of those alarms are less than critical.

The manual for the motherboard doesn't specify what the various "temp" readings are, I'm assuming maybe CPU0, Internal case, and CPU1, but that's just speculation on my part. At any rate, there is no way the range for a component would be "high = -17.0°C, hyst = -65.0°C) and I don't know how to reset that in sensord, or where ever.

I'm kind of stumped and don't know why this would be an intermittent thing.

Shrug.

Mark

The sensor min/max limits are in /etc/sensors.conf. If limits aren't specified, they seem to default to random values, hence the nonsensical values you are seeing.

68.0°C is higher than I would be comfortable with, but not overly alarming. If it were my system, I'd be looking for ways to improve the airflow. If there are any air filters, check they aren't clogged up with dust.

Have you run a memory test on the system? I had several memory DIMMs fail after 6-12 months of use, and the symptoms were intermittent faults just like you are seeing.

Ran Memtest86+ on it and dozed off. When I woke up, the machine had shut down.

I have a stick of RAM ordered from Crucial. I'll keep futzing with it in the meantime.

Thanks for the assist. I'll get back here if it doesn't resolve with new RAM.

Mark

I would try clearing the CMOS as well. Do you know how to do this?

daniewicz wrote:I would try clearing the CMOS as well. Do you know how to do this?

Not off the top of my head, but I do have the mobo manua.Will look at doing that when I get the chance.

OK. Let us know if you need some guidance.

Clearing the CMOS is always my first step when I have hardware issues.

Finally opened up the case last week and the news was not good.

The 80mm fan in the psu was in a bind and not turning. It was providing power because all the lights were buring, the case fans and the cpu cooler were turning as expected.

The fan on the video card was not turning and it wasn't producing signal. I just happen to have a spare GT8600 and slotted it in.

I wasn't getting a "no signal" on the monitor, but I wasn't seeing a POST either. Retting cmos and swapping memory did not help.

At this point, I'm on the verge of replacing the mobo with a refurb of the same make/model and either replaceing the 80mm fan in the psu, or springing for a new psu. I'm hoping that I won't have to replace a $50-$60 cpu.

Any other ideas out there?

Removing memory and graphics card at once, do you get a POST code, does it beep?

Damn.

Okay, I stripped out the RAM and video card and booted the machine. I got three long beeps -- memory missing.

I put the video card and the new stick of RAM in and turned it on, and it booted.

At this point, I have the new ram in DIMM 0 and the old stick in DIMM 1. It's not seeing the old RAM in DIMM 1. I suspect that stick is bad but I have no idea why it behaved the way it did. I do know that the 80mm fan in the psu is not working like it should. It's stiff and turning slower than expected. I'm not sure whether to try and replace that fan or just buy a new psu. In the meantime, I'll see if it starts falling over again and decide whether I should pull that old stick of RAM.

Opinions?

What make/model is that PSU?

Personally, if I suspected a PSU had been running overly hot for a while, I would play safe and replace it. A PSU failure can take a lot of other things with it!

The PSU is a Thermaltake W0093RU Silent Purepower TR2 500Watt PFC. And I agree that it's been run way hot and that's one of the issue I've been dealing with/am still dealing with.

I'm looking at another PSU. Something like this:Logisys PS550A-BK.

Anybody see anything wrong with it?

Edit: or maybe this one: http://www.zagall.com/product_p/2768.htm

IMHO saving on the PSU is saving at the wrong end. Better buy some PSU from a well-known brand which is reliable, like Enermax, Seasonic or similar.

doktor5000 wrote:IMHO saving on the PSU is saving at the wrong end. Better buy some PSU from a well-known brand which is reliable, like Enermax, Seasonic or similar.

I'll second that, a fire in the server room is not anything I'd want to face... again. I don't think the PSU that burned was really capable of it's rated capacity. (allegedly 500 watts) Dang good thing I happened to be present when it finally flared up. (the 'server room' in question happens to be in my home)

I wouldn't scrimp on a power supply.

Okay, so: http://www.compsource.com/ttechnote.asp ... 110&src=PW

I'm going to snag a cheap replacement locally until I can get this in. My HTPC has been down for a month now and I want my box back.

Okay, snagged a 500W Coolermaster at my local PC shop. It should hold me unitl I can get some money together. This has resolved the issue, so to recap:

nVidia 8600GT video card - fan failed
SuperTalent DDR2 x 1, 1GB RAM - failed
500W Thermaltake PSU - fan failed (heat rendered PSU unreliable)

This equipment has essentially been up and running 24/7 for more than 5 years. Not bad for consumer hardware in my opinion.

I'm marking this one resolved.

Thanks.


Mark