[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#843349: linux-image-4.7.0-1-amd64: Intermittently high system load, sluggish response, no /proc/<pid>/stat utime,stime,etc. reported



On Sun, 13 Nov 2016 04:23:54 +0000 Ben Hutchings <ben@decadent.org.uk> wrote:
> Control: tag -1 moreinfo
>
> On Sun, 2016-11-06 at 00:02 -0700, Tom Lee wrote:
> [...]
> > Downgrading to linux-image-4.6.0-1-amd64 fixed all symptoms. Haven't yet
> > tried 4.8.0-1 from unstable, not sure if it's impacted.
> [...]
>
> Please do.
>
> Ben.
>
> --
> Ben Hutchings
> Nothing is ever a complete failure; it can always serve as a bad
> example.
>

Hi,

I stumbled on the same problem using sid's 4.8.0-1 kernel.
I had been using linux-image-4.7.0-1-amd64 for a while without
trouble, and noticed the issue after upgrading to 4.8.0-1.
I then reverted to using 4.7.0-1... and still the issue persisted!

A colleague of mine told me that he had noticed the problem as well a
while ago. He switched back to kernel version 4.6.0-1, which fixed it.
Later he started using 4.7.0-1 again, but the problem had gone away.
Nasty.

After much rebooting and hair-pulling, I noticed that we three had one
thing in common:

sys_vendor: Alienware
product_name: Aurora-R4
product_version: 00
chassis_vendor: Alienware
chassis_version: 00
bios_vendor: Alienware
bios_version: A11


So here's my theory: there's a bug at some point in the reboot process
that leaves the CPU in a weird state in which some counter are not
getting updated at all. This explains the utime/stime values being
always 0, and this would also probably wreak havoc on the scheduler if
it relied on these counters to account for time slices, explaining the
overall sluggishness.

Since it seems to only affect Aurora-R4 users, I'm guessing the
motherboard might be doing something funky to reset the CPU during
reboots. This behavior may have been revealed by changes in the
4.7.0-1 kernel. So, motherboard bug? Kernel regression? Both? I don't
know.

Long story short, I performed a cold reboot (even pulling the plug for
extra safety), and my system went back to running smoothly using sid's
linux 4.8.0-1.

Hope this helps,

Christophe Aguettaz


Reply to: