Scheduled Maintenance: We are aware of an issue with Google, AOL, and Yahoo services as email providers which are blocking new registrations. We are trying to fix the issue and we have several internal and external support tickets in process to resolve the issue. Please see: viewtopic.php?t=158230

 

 

 

Computer freezes and crashes.

If none of the specific sub-forums seem right for your thread, ask here.
Post Reply
Message
Author
Laserlab
Posts: 7
Joined: 2018-11-08 11:26

Computer freezes and crashes.

#1 Post by Laserlab »

Hello everyone,
I'm on Debian 9, i386, with xfce. I installed it yesterday using a live image, then run upgrades.
My computer is a laptop, a Toshiba Portégé R600-11B.
The computer randomly freezes when performing a task (e.g.: browsing the web with firefox, editing text with LibreOffice), the screen becomes black, then it shuts down and reboot by itself.

Actual problem:
When performing ram medium-consuming tasks, the fan starts to overrun very quickly, then the computer freezes, shuts down and reboot.
I put a cooler pad under the computer, figuring it was some overheating issue: not only it changed nothing but also I realised by touching it that the computer was not hot.

History of the problem:
I was previously on Lubuntu. Since two or three weeks ago, the computer would crash the same way, two times before the login panel (on the time between Grub and the login panel), then a third one once logged in (usually from 1 to 10 minutes after starting up). Then, it would work for hours without problem.
This was being very problematic. I reinstalled Lubuntu with a previous version (15.10) then run some updates, and it started crashing again, the same way.
Then I decided to install Debian. The network would not work so I had to add the package firmware-iwlwifi_20161130-3_all.deb using dpkg -i, but, despite configuring the mirrors on the source.list file and adding some software such as Hexchat or the LibreOffice suit, I did not changed the basics settings.

Here is the output of dmesg:
https://pastebin.com/JkNT5uKM

I'd like, please, some help to solve this.

EDIT: I ran a memory test and there was no error.

Segfault
Posts: 993
Joined: 2005-09-24 12:24
Has thanked: 5 times
Been thanked: 17 times

Re: Computer freezes and crashes.

#2 Post by Segfault »

Sounds like possible poor thermal contact between CPU and cooler. I'd refresh the thermal paste. While in there I'd also clean the dust out. With 2 GB of RAM or more you shouldn't use 32 bit OS.

Laserlab
Posts: 7
Joined: 2018-11-08 11:26

Re: Computer freezes and crashes.

#3 Post by Laserlab »

UPDATE: now it crashes on Debian on the time between Grub and the login panel, twice, as it did on Lubuntu.
Segfault wrote:Sounds like possible poor thermal contact between CPU and cooler. I'd refresh the thermal paste. While in there I'd also clean the dust out.
Ok, I'll try this then. Do you see a logic correlation between this and the Grub-to-login time interval ?
With 2 GB of RAM or more you shouldn't use 32 bit OS.
So I should switch to 64bits, but why ? Aren't 64bit aps more RAM-consuming (10-15%), and the memory limitation problem starting only with 4GB ? (I have less than 4, RAM+graphical card joined).

User avatar
bw123
Posts: 4015
Joined: 2011-05-09 06:02
Has thanked: 1 time
Been thanked: 28 times

Re: Computer freezes and crashes.

#4 Post by bw123 »

I think the suggestion about the heatsink is good. Make sure all (4?) clamps are in place, nothing interfering at all or touching the fan. You just need a tiny dab of paste, clean the old off first if you have removed and reset the cooler.

Did you select pae kernel? Why, if you have less than 4GB of ram? also what was the iommui= stuff in /proc/cmdline for? That's a nice cpu, and 3 gigs should be plenty, why not use amd64?
resigned by AI ChatGPT

Segfault
Posts: 993
Joined: 2005-09-24 12:24
Has thanked: 5 times
Been thanked: 17 times

Re: Computer freezes and crashes.

#5 Post by Segfault »

RE: 32 bit
First, the problems start from 1-2 GB, as explained here.
Second, while 32 bit still has limited usage in special cases it is no good for desktop use. Support for 32 bit has been phased out. FOSS still works, but proprietary software does not, Skype for instance. And Chrome. The 32 bit era is over. It is year 2018.

Edit: And memtest can only tell you when the RAM is bad, it won't tell you if it is good. Meaning, it may take several runs before it finds a bad region. I recall a report someone let it run for 7 days and then memtest reported bad RAM.

User avatar
debiman
Posts: 3063
Joined: 2013-03-12 07:18

Re: Computer freezes and crashes.

#6 Post by debiman »

the laptop is almost 10 years old, so checking for dust in the fan or hardware fatigue does make sense.

in your dmesg i searched for "error" and came up with this:

Code: Select all

[    0.844177] tpm_tis 00:06: 1.2 TPM (device-id 0xB, rev-id 16)
[    0.924230] tpm tpm0: [Hardware Error]: Adjusting reported timeouts: A 750->750000us B 2000->2000000us C 750->750000us D 750->750000us
[    1.004217] tpm tpm0: Adjusting TPM timeout parameters.
[    1.084225] tpm tpm0: [Hardware Error]: Adjusting reported timeouts: A 750->750000us B 2000->2000000us C 750->750000us D 750->750000us
[    1.164214] tpm tpm0: Adjusting TPM timeout parameters.
[    1.408021] clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x1421470e227, max_idle_ns: 440795242981 ns
[    2.044119] tpm tpm0: A TPM error (7) occurred attempting to read a pcr value
[    2.044184] tpm tpm0: TPM is disabled/deactivated (0x7)
i have no idea what it means, please do some searches+occurred+attempting+to+read+a+pcr+value).

Laserlab
Posts: 7
Joined: 2018-11-08 11:26

Re: Computer freezes and crashes.

#7 Post by Laserlab »

Hello again,
Segfault wrote:RE: 32 bit
First, the problems start from 1-2 GB, as explained here.
Second, while 32 bit still has limited usage in special cases it is no good for desktop use. Support for 32 bit has been phased out. FOSS still works, but proprietary software does not, Skype for instance. And Chrome. The 32 bit era is over. It is year 2018.

Edit: And memtest can only tell you when the RAM is bad, it won't tell you if it is good. Meaning, it may take several runs before it finds a bad region. I recall a report someone let it run for 7 days and then memtest reported bad RAM.
Ok that's some interesting information on there. Funny thing, this 7-days memtest. I won't take the report too seriously then.

Laserlab
Posts: 7
Joined: 2018-11-08 11:26

Re: Computer freezes and crashes.

#8 Post by Laserlab »

bw123 wrote:I think the suggestion about the heatsink is good. Make sure all (4?) clamps are in place, nothing interfering at all or touching the fan. You just need a tiny dab of paste, clean the old off first if you have removed and reset the cooler.

Did you select pae kernel? Why, if you have less than 4GB of ram? also what was the iommui= stuff in /proc/cmdline for? That's a nice cpu, and 3 gigs should be plenty, why not use amd64?
Heatsink:
I agree too with the heatsink possibility, it will just take me more time to set it in order (getting the paste, etc).

IOMMUI:

Before asking for help here, I first searched by myself. I learned that I could check the dmesg output. I found some errors, started with a drm one, looked for solutions. I don't remember what was the error exactly, but the possible solution was similar to that one:
https://askubuntu.com/questions/1025574 ... re-problem
It fixed the error, but not the crash thing, so I decided to remove the fix, updated grub again, reboot, and get a dmesg output only after that so there would be the less "user noise" possible when I exposed the problem here.

PAE, architecture:
There was no obvious "pae" choice when I installed debian. I'm just learning about PAE, now that it has been mentionned in this conversation. Furthermore, the information I had on 64bits VS 32bits seems to be unadequate or updated. With the information I had (not sure if necessary to detail it) I was looking for a 32 bits system, usually presented in opposition with 64bits for what I previously saw. I went on the Debian website.
-> Getting Debian
-> Complete installation image
-> Using HTTP
-> Stable
And then "CD suitable" section.
Here are the choices:
amd64, arm64, armel,armhf, i386, mips, mips64el, mipsel, ppc64el, s390x, source, multi-arch
I was not expecting these terms. I knew the i686 one as an equivalent on 32 bits, so searched on the net if i386 could mean "32 bits", most of the answers available were positive. I downloaded this i386 option then. Apparently i386 is also the default choice for 32-bits for CDs, on the first page, "Getting Debian".
Once I finished downloading the file, there was no "PAE" terms in the .iso neither.

Laserlab
Posts: 7
Joined: 2018-11-08 11:26

Re: Computer freezes and crashes.

#9 Post by Laserlab »

debiman wrote:the laptop is almost 10 years old, so checking for dust in the fan or hardware fatigue does make sense.

in your dmesg i searched for "error" and came up with this:

Code: Select all

[    0.844177] tpm_tis 00:06: 1.2 TPM (device-id 0xB, rev-id 16)
[    0.924230] tpm tpm0: [Hardware Error]: Adjusting reported timeouts: A 750->750000us B 2000->2000000us C 750->750000us D 750->750000us
[    1.004217] tpm tpm0: Adjusting TPM timeout parameters.
[    1.084225] tpm tpm0: [Hardware Error]: Adjusting reported timeouts: A 750->750000us B 2000->2000000us C 750->750000us D 750->750000us
[    1.164214] tpm tpm0: Adjusting TPM timeout parameters.
[    1.408021] clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x1421470e227, max_idle_ns: 440795242981 ns
[    2.044119] tpm tpm0: A TPM error (7) occurred attempting to read a pcr value
[    2.044184] tpm tpm0: TPM is disabled/deactivated (0x7)
i have no idea what it means, please do some searches+occurred+attempting+to+read+a+pcr+value).
I forgot to mention that I already checked that one too. My bad.
In the solution I found, I had to check if using

Code: Select all

blkid -o list
would present two or more UUID's being the same, and that was not my case.

EDIT:
I also tried this solution:
http://forums.debian.net/viewtopic.php? ... sh#p683589
since it presents similarities with my case (Toshiba, Stretch, etc), but that did not fixed my problem.

shep
Posts: 423
Joined: 2011-03-15 15:22

Re: Computer freezes and crashes.

#10 Post by shep »

Along the lines of overheating/thermal zone shutdown, consider checking the BIOS. Many provide hardware monitoring where fan rpm and temperatures are reported.
Some laptops have more than 1 fan. Most fans eventually fail.

I would also web search your make/model. If you're lucky, the disassembly is documented and you can see what it would entail to clean the lint out and replace thermal paste/fans etc.

Laserlab
Posts: 7
Joined: 2018-11-08 11:26

Re: Computer freezes and crashes.

#11 Post by Laserlab »

Hello,

the thermal paste has been changed, and everything cleaned.
The computer still crashes.
I'll underline the fact that it crashes mainly at startup, usually twice, but rarely after 2-3 hours of utilisation.

@shep: I checked but it seems that there is no hardware monitoring in the BIOS.

What else can I check/debug ?

Segfault
Posts: 993
Joined: 2005-09-24 12:24
Has thanked: 5 times
Been thanked: 17 times

Re: Computer freezes and crashes.

#12 Post by Segfault »

Typical for dying capacitor. Did you notice any bulging electrolytic caps when you had it open?

Laserlab
Posts: 7
Joined: 2018-11-08 11:26

Re: Computer freezes and crashes.

#13 Post by Laserlab »

Segfault wrote:Typical for dying capacitor. Did you notice any bulging electrolytic caps when you had it open?
That's possible. I did not paid attention to it ! I'll open the machine and check for this especifically.

UPDATES ON TOPIC:
Some things have changed since the thermal paste has been changed and the dust cleaned:
-There are still reboots at starting-up, most especifically on the time between I select 'Debian' on GRUB and the 20 first seconds on Desktop.
-The reboots when in use have stopped, I can launch and use any application without crashes nor freezes.

I then suppose that my issue is/was a multiple issue, either hardware-hardware issue, either hardware-software issue.
In the case of hardware-hardware issue, I'm keeping the suggestion of Segfault as top one (so thermal paste+dying capacitor).
In the case of hardware-software issue, I don't know. I started checking for the errors provided by

Code: Select all

journalct -p err
but it seems that they are not related, for what I actually saw. Elsewhere, ideas are wellcome.

Post Reply