docs: limitations: document 'unhandled rdmsr' dmesg warnings #884
Conversation
|
kubernetes qa-passed 👍 |
|
Hmm, some odd test timeout on the F26 CI - @chavafg - something we've seen and know about?? |
docs/limitations.md
Outdated
| @@ -33,6 +33,8 @@ | |||
| * [runtime commands](#runtime-commands) | |||
| * [`init` command](#init-command) | |||
| * [`spec` command](#spec-command) | |||
| * [Other notes](#other-notes) | |||
| * [Host `rdmsr` warnings](#host-rdmsr-warnings) | |||
|
|
|||
| As Intel® Clear Containers utilises Virtual Machines (VM) to enhance | |||
There was a problem hiding this comment.
@klynnrif - that change is not really part of this PR - it is unrelated (already existed). No doubt it emanated from the English'ness of @jodh-intel and myself :-)
I'll add that as an add-on fix to this PR - but we should be very aware of reviewing and feeding back information generally only pertinent to the PR in question - if other major things pop out to you then it is probably best to open a new Issue to request them to be fixed.
There was a problem hiding this comment.
Yes, the "s" variant is much more common I think over here.
At the risk of starting a language war... imho if the word is in the OED (and both spellings are listed for this word), we should not try to Americanise it (or should that be Americanize? 😄) and consider accepting both spellings. However, there is something to be said for consistency. Looking towards kata, I think we can avoid a lot of this confusion going forward if we define our approach somewhere (probably https://github.com/kata-containers/documentation (either wiki or in-tree)).
See the related issue I raised way back on this: intel/cc-oci-runtime#788.
There was a problem hiding this comment.
@grahamwhaley @jodh-intel Thanks for the explanation :) I figured that was the case and did question myself before including it. I agree with @jodh-intel on no war and consistency :) Thanks!
docs/limitations.md
Outdated
| ## Other notes | ||
|
|
||
| This section contains other useful information realated to limitations, | ||
| warnings etc. to aid in understanding and diagnosis of what is a known |
There was a problem hiding this comment.
Adding comma before "etc.": warnings, etc. to aid in understanding...
docs/limitations.md
Outdated
|
|
||
| ### Host `rdmsr` warnings | ||
|
|
||
| When running Clear Containers, you may witness some warnings in the host |
There was a problem hiding this comment.
When you run Clear Containers, you might see some warnings in the host
docs/limitations.md
Outdated
| by `QEMU/KVM`. In the case of Clear Containers, these warnings are | ||
| normally related to unsupported perf event counters. | ||
|
|
||
| These are in fact benign warnings. Unfortunately by default they appear |
There was a problem hiding this comment.
By default, these benign warnings appear as 'red warnings', which can be alarming to the end users and system administrators, in the dmesg logs.
docs/limitations.md
Outdated
|
|
||
| https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/x86/kvm/x86.c?id=fab0aa3b776f0a3af1db1f50e04f1884015f9082 | ||
|
|
||
| It may take some time for those changes to be implemented in |
We get warnings in the host dmesg such as: ``` [ 319.406575] kvm [2415]: vcpu0 unhandled rdmsr: 0x1c9 [ 319.483944] kvm [2415]: vcpu0 unhandled rdmsr: 0x64e [ 319.483966] kvm [2415]: vcpu0 unhandled rdmsr: 0x34 ``` that although benign, can understandably perturb users. Document them so at least information about the warnings can be found without having to submit a new issue etc. Fixes: clearcontainers#883 Signed-off-by: Graham whaley <[email protected]>
Americanise utlise -> utilize. Signed-off-by: Graham whaley <[email protected]>
grahamwhaley
force-pushed
the
20180102_rdmsr_warnings
branch
from
d2a4e71
to
e972f27
Compare
|
Fixes should now be applied @klynnrif - thanks! |
|
kubernetes qa-passed 👍 |
|
Doc-only PR, so ignoring the F26 failure (captured on clearcontainers/jenkins#32). |
Align with other network models. Fixes clearcontainers#884. Signed-off-by: Ruidong Cao <[email protected]>
We get warnings in the host dmesg such as:
that although benign, can understandably perturb users.
Document them so at least information about the warnings
can be found without having to submit a new issue etc.
Fixes: #883
Signed-off-by: Graham whaley [email protected]