12:12 p.m.
Heya,
This is a brief summary of discussions re `no_timer_check` Kernel
parameter from different bug trackers (see references below for more
details)
What?
-----
Make 'no_timer_check' Kernel parameter default for Fedora cloud images.
Why?
----
It has been reported (recently and in the past) many times[1][2][3] that
Kernel timer checks are not reliable/helpful in the context of
virtualization and it's useful to disable
This is also useful to elimiate false-positives in upstream
OpenStack CI testing[1][2].
Notes
-----
- 'no_timer_check' parameter is added implicitly when the kernel
detects it is booting under KVM.
However when booting under QEMU TCG (software emulation) you have to
add this to the guest kernel command line explicitly.
(Thanks Richard WM Jones.)
- About 'no_timer_check', from Kernel documentation[4]:
'no_timer_check' [X86,APIC] Disables the code which tests for
broken timer IRQ sources.
References
----------
[1] https://bugs.launchpad.net/cirros/+bug/1312199 -- cirros 0.3.1
fails to boot
- Relevant comments: #1, #2, and #4
[2] https://bugzilla.redhat.com/show_bug.cgi?id=1102592 -- F20 could
image affected by timing issues when using with soft-qemu
[3] https://bugzilla.redhat.com/show_bug.cgi?id=502058#c15 -- qemu
-no-kvm guest hangs at during timer setup; works with noapic
[4] https://www.kernel.org/doc/Documentation/kernel-parameters.txt
--
/kashyap
12:16 p.m.
On Wed, Jun 18, 2014 at 10:42:51PM +0530, Kashyap Chamarthy wrote:
Heya,
This is a brief summary of discussions re `no_timer_check` Kernel
parameter from different bug trackers (see references below for more
details)
What?
-----
Make 'no_timer_check' Kernel parameter default for Fedora cloud images.
Why?
----
It has been reported (recently and in the past) many times[1][2][3] that
Kernel timer checks are not reliable/helpful in the context of
virtualization and it's useful to disable
This is also useful to elimiate false-positives in upstream
OpenStack CI testing[1][2].
Notes
-----
- 'no_timer_check' parameter is added implicitly when the kernel
detects it is booting under KVM.
However when booting under QEMU TCG (software emulation) you have to
add this to the guest kernel command line explicitly.
FYI, upstream OpenStack now explicitly adds this kernel parameter when
possible, but it can only do this when booting from a kernel/initrd.
If the guest is booting from a virtual BIOS, there's no option but
to add it to the guest image grub.conf manually.
commit 6b86a61fee15ce1237303fab2f7896f8c3bcad47
Author: Attila Fazekas <afazekas(a)redhat.com>
Date: Wed May 28 09:19:29 2014 +0200
Use no_timer_check with soft-qemu
The Linux kernel timer check not working properly
when the hypervisor's thread preempted by the host CPU scheduler.
The timer check is automatically disabled with other types
of hypervisors including the hardware accelerated kvm,
but timer_check is not disabled when qemu used without hardware acceleration.
This issue is frequently mischaracterized as an SSH connectivity issue and
causes rechecks and occasional boot failures.
This change adds no_timer_check kernel parameter when we are using
uec images with qemu.
Closes-Bug: #1312199
Change-Id: I3cfdfe9048fe219fc12cdac8a399b496f237e55e
Regards,
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
12:23 p.m.
On Wed, Jun 18, 2014 at 06:16:04PM +0100, Daniel P. Berrange wrote:
On Wed, Jun 18, 2014 at 10:42:51PM +0530, Kashyap Chamarthy wrote:
[. . .]
> Notes
> -----
>
> - 'no_timer_check' parameter is added implicitly when the kernel
> detects it is booting under KVM.
>
> However when booting under QEMU TCG (software emulation) you have to
> add this to the guest kernel command line explicitly.
FYI, upstream OpenStack now explicitly adds this kernel parameter when
possible, but it can only do this when booting from a kernel/initrd.
If the guest is booting from a virtual BIOS, there's no option but
to add it to the guest image grub.conf manually.
Yup, noticed this fix from LP bug -- 1312199. I see exactly your
point (it works only when you use '--append' w/ external
Kernel/initrd) echoed in comment #14[1], by rjones.
[1] https://bugs.launchpad.net/cirros/+bug/1312199
commit 6b86a61fee15ce1237303fab2f7896f8c3bcad47
Author: Attila Fazekas <afazekas(a)redhat.com>
Date: Wed May 28 09:19:29 2014 +0200
Use no_timer_check with soft-qemu
The Linux kernel timer check not working properly
when the hypervisor's thread preempted by the host CPU scheduler.
The timer check is automatically disabled with other types
of hypervisors including the hardware accelerated kvm,
but timer_check is not disabled when qemu used without hardware acceleration.
This issue is frequently mischaracterized as an SSH connectivity issue and
causes rechecks and occasional boot failures.
This change adds no_timer_check kernel parameter when we are using
uec images with qemu.
Closes-Bug: #1312199
Change-Id: I3cfdfe9048fe219fc12cdac8a399b496f237e55e
--
/kashyap
3594
days inactive
3594
days old
2 comments
2 participants
participants (2)
-
Daniel P. Berrange -
Kashyap Chamarthy