Upgrade Kernel; or What?

r00ster · 07-22-2012, 12:13 PM

I've been using Linux as a Home Desktop for a few years now; primarily because I prefer FOSS philosophy and I don't like MS "for profit" practices and procedures.

Same time, I have to concede Linux takes more of my time; i.e., has a steeper and longer learning curve than either Windows or Mac. When I upgraded from Etch to Squeeze, I ended up with problems I didn't, and couldn't have, anticipated and which I am still trying to sort out when I can make the time. What I want to say here is that for the last couple of years, I spend far more time maintaining/upgrading/fixing the 'System' than I do 'End Using' it.

I still haven't got 6.0 (Squeeze) to stop freezing 3-4-5 + times a session and now Wheezy is breathing over my shoulder.

I would appreciate some insight into what is meant by some recommendations I'm seeing that say 'we' should be upgrading to Linux Kernel 2.6.39? AFAICT, I'm running 2.6.32-5-686 and I've been doing the updates and upgrades as soon as the notifications appear.

So, what up with upgrading the kernel?

rooster

snowday · 07-22-2012, 12:26 PM

Hi r00ster, "freezing 2-4-5 + times a session" is NOT normal behavior for Debian Stable! Debian is renowned for its stability, however, it is possible that certain combinations of hardware and software might cause problems. I am stunned that you stuck with it for as long as you did if it is freezing up that often.

What I would do is burn a few Live CD/USB's and find a distro that works well on your hardware. The vast majority of Linux users have zero freezes, and Linux is the OS of choice on servers that must have 100% uptime. You deserve a 100% stable and bug-free experience with Linux, keep searching until you find it!

To answer your specific question: It can't hurt to install a newer kernel in your Squeeze to see if it solves the problem. How could it be any worse than what you have now, right? The easiest way to do this is using backports, and the current backports kernel is 3.2. Another suggestion is that Wheezy is in "freeze" and is considered "stable enough" for most users. You can try a Live CD of Wheezy and if it solves your freezing problems, abandon Squeeze and go for a fresh reinstall of Wheezy. Other distros that are known to be very stable include CentOS and Slackware.

If you want to stick with Squeeze and troubleshoot the problem (I wouldn't, personally--I would just do a fresh reinstall of something that works hassle-free) then a good place to start is by telling us about your hardware.

jefro · 07-22-2012, 02:45 PM

Almost no machine is tested at the factory with linux. So the freezing may or may not be cured. I agree that you may need to try some distro's or even start with bios settings to slow it down, check or get best drive cables, run memtest and or hard drive diags for a day or so, watch temps and voltages too.

Testing a different kernel can't hurt but you don't see errata in kernel these days for that issue.

I agree that live cd's are a great way to see if some disto could work. Might try Gentoo. Dunno what they did but it usually works.

You are right is the steep learning curve and taking time.

r00ster · 07-23-2012, 07:36 AM

First, thanks y'all for your understanding and informative and supportive replies.

snowpine & jefro;

In an ideal world (mine, that is) I would appreciate learning how to diagnose this specific problem. I upgraded from Etch (which worked like a charm) to Lenny. But freezes and other errata made me deep-six Lennhy before activating my local providers email accounts or loading my documents & settings from backup media.

Here's my unit's profile.

Quote:

CPU: Intel Celeron D Processor 330 (2.66GHz, 533MHz FSB, 256KB L2 cache)

Operating Systems
Master HDD
Debian GNU/Linux 4.0, Etch: Kernel: linux-image-2.6.18-4-686
Debian GNU/Linux 6.0, Squeeze: Kernel: linux image 2.6.32-5-686

Chainloaded Slave HD
Genuine Microsoft Windows XP Home

Chipset: Intel 845GV chipset

Memory: 512MB DDR (PC2700); expandable to 2GB

HDD (Master): Maxtor DiamondMax Plus 9 EIDE 160GB ATA-133
HDD (Slave): WD Caviar WD800 BB-JHAD 80GB (OEM XP-Home)

Optical Drive: CD-RW/DVD Combo Drive
ODD (Slave): Samsung SH-S182 DVD+RW

Media Reader: 8-in-1 digital media manager (Secure Digital (SD), Smart Media,
Compact Flash, Micro Drive, Memory Stick, Memory Stick PRO,
Multimedia Card, USB 2.0

Video: Intel Extreme Graphics 3D 64MB DDR Shared memory

Sound: AC '97 audio

Network: Intel PRO 10/100Mbps integrated Ethernet LAN

Modem: 56K ITU v.92-ready fax/modem

Peripherals:
Standard emachines multi-media keyboard
Logitech Fast RF Cordless Mouse (LX5- S'Ware: Logitech Setpoint 2.4a)
Monitor: Dell CRT 16 in.
Labtec amplified stereo speakers and subwoofer

Interfaces: 5 USB 2.0 ports (1 in media manager, 4 in back),
VGA external connector, 1 Serial, 1 Parallel, 2

DRIVER(S)
rooster@royrogers:~$ cat /var/log/Xorg.0.log | grep -i driver
X.Org Video Driver: 6.0
X.Org XInput driver : 7.0
<snip>
(II) intel: Driver for Intel Integrated Graphics Chipsets: i810,
(II) VESA: driver for VESA chipsets: vesa
(II) FBDEV: driver for framebuffer: fbdev
ABI class: X.Org Video Driver, version 6.0 <snip>

# apt-cache policy xserver-xorg-video-intel
xserver-xorg-video-intel:
Installed: 2:2.13.0-6
Candidate: 2:2.13.0-6
Version table:
2:2.15.0-3~bpo60+2 0
100 http://backports.debian.org/debian-backports/ squeeze-backports/m86

I've done occasional forum and literature reviews on this freezing problem for more than a year. There seems to be a handful of desktoppers across the spice rack of 'flavours' with this persistent problem. So far I haven't been fortunate enough to find a diagnosis or sol'n that is clear-cut; ...just tentative references to myriad possibilities, some of which y'all have noted above.

For the benefit of my understanding, what is the significant difference between a 'kernel version' and a 'kernel upgrade'? I read Linus's release notes and I see what he purposed to change. But I am pretty limp-headed as to why these aren't part-and-parcel of the 'update/upgrade' prompted regime available from my repo's. It's naive I know, but it seems to me that if you 'upgrade' the kernel, you'd be running a different version; albeit the same Distribution. Semantics can be a ....

Can/should a 'kernel upgrade' be accomplished from terminal? Or is it necessary to d/l it from a web site? (I know to back up files beforehand) I do have backports in my 'sources.list'.

Code:

<snip> # to update iceweasel for gmail browser requirements - per:http://mozilla.debian.net/
deb http://backports.debian.org/debian-backports squeeze-backports main
deb http://backports.debian.org/debian-backports squeeze-backports main contrib non-free
deb http://mozilla.debian.net/ squeeze-backports iceweasel-release <snip>

WRT the freeze issue, before tinkering with the kernel, might posting the results of some log or other provide a clue? I don't want to exploit forum bandwidth if the P/Os are not relevant. I'm not savvy enough to interpret them. E.g.,

Code:

# tail -f /var/log/messages 
# less /var/log/messages 
# more -f /var/log/messages 
# vi /var/log/messages

rod

cynwulf · 07-23-2012, 08:06 AM

Quote:

Chipset: Intel 845GV chipset

It's likely to be a problem with the intel driver.

Etch and Lenny used the i810 driver for the 8xx chips, whereas squeeze (and later) uses the i915 driver.

You should first upgrade the kernel, the xorg driver and drm stacks from backports and hope that resolves it. If you search both here and at forums.debian.net, you'll find plenty of threads relating to this.

Try:

Code:

$ grep i915 /var/log/messages

And see if there are any useful errors which may confirm this...

r00ster · 07-23-2012, 09:08 AM

By your command:

Code:

# grep i915 /var/log/messages                      
Jul 22 02:35:47 royrogers kernel: [    9.393705] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
Jul 22 02:35:47 royrogers kernel: [    9.680560] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
Jul 22 03:51:12 royrogers kernel: [    8.349625] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
Jul 22 03:51:12 royrogers kernel: [    8.638618] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
Jul 22 04:23:47 royrogers kernel: [    8.395288] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
Jul 22 04:23:47 royrogers kernel: [    8.685769] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
Jul 22 04:41:55 royrogers kernel: [    8.696931] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
Jul 22 04:41:55 royrogers kernel: [    9.392941] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
Jul 22 04:59:56 royrogers kernel: [    7.799598] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
Jul 22 04:59:56 royrogers kernel: [    8.085189] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
Jul 22 05:59:11 royrogers kernel: [   10.489568] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
Jul 22 05:59:11 royrogers kernel: [   10.756177] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0
Jul 22 07:26:58 royrogers kernel: [    8.519133] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
Jul 22 07:26:58 royrogers kernel: [    8.769469] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

I don't think that P/O goes back to the last freeze. It's just my current session.

rod

cynwulf · 07-25-2012, 05:34 AM

grep dmesg.0 instead.

r00ster · 07-26-2012, 09:53 AM

Quote:

Originally Posted by caravel

grep dmesg.0 instead.

No output for:

Code:

root@royrogers:/home/rooster# grep dmesg.0

Was I supposed to assume you intended ?:

Code:

root@royrogers:/home/rooster# grep i915 /var/log/dmesg.0
[    5.672458] i915 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[    5.672468] i915 0000:00:02.0: setting latency timer to 64
[    5.941122] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

rod