Spontaneous rebooting

[Kelly Dean]

Am I the only one having a problem with Qubes spontaneously rebooting on Intel hardware? Only other reports I see are about AMD problems, but I'm using an Intel Core i3.

Happens every few weeks. Sometimes just 1 or 2 weeks, sometimes 5 or 6. Got it on Qubes 3.2, and now 4.0 too (new installation, not upgrade), multiple times.

Unlikely to be a hardware problem. The system passed both memtest86 and a multi-day mersenne prime stress test. And other OSes tested on this hardware before I switched to Qubes, including Debian and Windows, never had a problem.

The rebooting seems completely random. No apparent trigger, and no warning. Acts like an instant hard reset. Sometimes even when the system is idle, and I haven't touched the console for hours.

It's wearingly inevitable enough that I don't even bother intentionally rebooting after system updates anymore, in order to minimize how many reboots I have to deal with (setting my workspace back up is an ordeal), because I know the system will end up spontaneously rebooting a week or two later anyway.

[Michael Siepmann]

I'm having this problem too. I hadn't had it for a while but in the past week or so it's happened a few times. I have a Lenovo ThinkPad T440p with Intel Core i7, and Qubes 4.0 which I keep updated.

[David Hobach]

Same here, but only since 4.0 and since coreboot & ME-cleaner on a T530.

I've always suspected that it's related to memory kills (there was an
OOM issue on github), but there's absolutely nothing in the journal
after such a "forced" reboot.

[unman]

I dont see this on any machine, including long running desktops.
Is it possible that you are suffering from over-heating? That would
account for symptoms.

[Chris]

Hello Michael,

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐

I had this problem which was due to Intel AMT Control being enabled in the BIOS. Since turning this off my system has not rebooted.

Regards,

Chris

-

Chris Willard

ch...@meliser.co.uk

Sent with ProtonMail Secure Email.

[brenda...@gmail.com]

There are some discussions in qubes-issues on github about torbrowser causing 100% cpu while idle, yet appearing to mostly work ok. Running a couple VMs with that bug might cause an overheat reboot on some systems...

[David Hobach]

On 4/13/19 8:29 PM, brenda...@gmail.com wrote:
> There are some discussions in qubes-issues on github about torbrowser causing 100% cpu while idle, yet appearing to mostly work ok. Running a couple VMs with that bug might cause an overheat reboot on some systems...

No Intel AMT (coreboot), no overheating (~50-60 degrees when idle), no
torbrowser here... But thanks for the hints anyway, maybe it'll help
someone!

I was referring to [1] earlier on. I didn't observe any log entries though.

I'll keep you posted if I ever find the root cause.

[1] https://github.com/QubesOS/qubes-issues/issues/3079

[Michael Siepmann]

Thanks for the suggestion, Chris. I appreciate it. I checked and Intel AMT was already disabled so I guess something else must be causing it in my case.

[Michael Siepmann]

I'm now monitoring temperatures with the "sensors" command in a dom0
terminal and although the problem hasn't yet happened again, the
temperatures I'm seeing are often getting close to or even reaching
"Critical":

coretemp-isa-0000
Adapter: ISA adapter
Package id 0: +100.0°C  (high = +84.0°C, crit = +100.0°C)
Core 0:        +99.0°C  (high = +84.0°C, crit = +100.0°C)
Core 1:       +100.0°C  (high = +84.0°C, crit = +100.0°C)

acpitz-virtual-0
Adapter: Virtual device
temp1:        +99.0°C  (crit = +200.0°C)

thinkpad-isa-0000
Adapter: ISA adapter
fan1:        4788 RPM

If over-heating is the cause, does that suggest a hardware problem, or
is there something I can do to configure Qubes differently to stop it
from getting so hot?

Thanks!

[David Hobach]

On 4/18/19 12:52 AM, Michael Siepmann wrote:
>> I dont see this on any machine, including long running desktops.
>> Is it possible that you are suffering from over-heating? That would
>> account for symptoms.
>
> I'm now monitoring temperatures with the "sensors" command in a dom0
> terminal and although the problem hasn't yet happened again, the
> temperatures I'm seeing are often getting close to or even reaching
> "Critical":
>
> coretemp-isa-0000
> Adapter: ISA adapter
> Package id 0: +100.0°C  (high = +84.0°C, crit = +100.0°C)
> Core 0:        +99.0°C  (high = +84.0°C, crit = +100.0°C)
> Core 1:       +100.0°C  (high = +84.0°C, crit = +100.0°C)
>
> acpitz-virtual-0
> Adapter: Virtual device
> temp1:        +99.0°C  (crit = +200.0°C)
>
> thinkpad-isa-0000
> Adapter: ISA adapter
> fan1:        4788 RPM
>
> If over-heating is the cause, does that suggest a hardware problem, or
> is there something I can do to configure Qubes differently to stop it
> from getting so hot?
>

That's not good. Usually CPUs shut down and cause a reboot as you
observed at 100-110 Celsius (CPUs may die at higher temperatures and
thus perform an emergency shutdown).

Looks like you have a hardware issue with your CPU fan. Either cleaning
the CPU fan and replacing the thermal paste or replacing the entire CPU
fan is likely to solve your issue. Youtube videos are your best friend
there.

I had similar syptoms in the past, replaced the fan and thermal paste
(apparently the seller had applied way too much) and went down to 70-80
Celsius at maximum load.

[Michael Siepmann]

Thank you! I will try cleaning it first and then replace it if needed. I
have some thermal paste that came with my Raspberry Pi.

[Jon deps]

On 4/13/19 8:01 PM, Michael Siepmann wrote:
>
>
> On 4/13/19 12:28 PM, Chris wrote:
>> Hello Michael,
>>
>> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
>> On Friday, April 12, 2019 3:24 PM, Michael Siepmann

>> <MS-CFHETcdP3ofiyS...@public.gmane.org> wrote:
>>
>>> On 8/10/18 12:37 PM, Kelly Dean wrote:
>>>
>>>> Am I the only one having a problem with Qubes spontaneously rebooting on Intel hardware? Only other reports I see are about AMD problems, but I'm using an Intel Core i3.
>>>>
>>>> Happens every few weeks. Sometimes just 1 or 2 weeks, sometimes 5 or 6. Got it on Qubes 3.2, and now 4.0 too (new installation, not upgrade), multiple times.
>>>>
>>>> Unlikely to be a hardware problem. The system passed both memtest86 and a multi-day mersenne prime stress test. And other OSes tested on this hardware before I switched to Qubes, including Debian and Windows, never had a problem.
>>>>
>>>> The rebooting seems completely random. No apparent trigger, and no warning. Acts like an instant hard reset. Sometimes even when the system is idle, and I haven't touched the console for hours.
>>>>
>>>> It's wearingly inevitable enough that I don't even bother intentionally rebooting after system updates anymore, in order to minimize how many reboots I have to deal with (setting my workspace back up is an ordeal), because I know the system will end up spontaneously rebooting a week or two later anyway.
>>>>
>>> I'm having this problem too. I hadn't had it for a while but in the
>>> past week or so it's happened a few times. I have a Lenovo ThinkPad
>>> T440p with Intel Core i7, and Qubes 4.0 which I keep updated.
>>>
>>
>> I had this problem which was due to Intel AMT Control being enabled in
>> the BIOS. Since turning this off my system has not rebooted.
>>
>> Regards,
>>
>> Chris
>>
>> -
>> Chris Willard

>> chris-HOuDeAah...@public.gmane.org <mailto:chris-HOuDeAah...@public.gmane.org>

>
>
> Thanks for the suggestion, Chris. I appreciate it. I checked and Intel
> AMT was already disabled so I guess something else must be causing it in
> my case.
>
>

and what does
#journalctl -r say if anything ?

[Michael Siepmann]

In a dom0 terminal if I type "journalctl -r" I see a lot of info.
Anything in particular I should look for?

[goo...@subvertising.org]

for instance if your machine spontaneously rebooted sometime between
2019-05-13 23:30 and 2019-05-14 you could issue:

journalctl --since "2019-05-13 23:30" --until "2019-05-14" -p debug
--no-pager | grep Reboot --before-context 42

this will output the final 42 log entries just before your machine
rebooted.
i'm stuck with the same issue (as i described here
https://www.mail-archive.com/qubes...@googlegroups.com/msg28344.html
) and my guess is you won't find anything useful in logs. anyway, i feel
your pain. calling unman et al. : what systems are sporting 60+ days
uptime?

[goo...@subvertising.org]

On 2019-05-14 11:33, goo...@subvertising.org wrote:
> calling unman et al. : what systems are sporting 60+
> days uptime?

according to unman, the i7 X230 does not reboot spontaneously.