kvm tuning guide

kvm tuning guide

[Avi Kivity]

I wrote a short tuning guide for kvm, 
http://www.linux-kvm.org/page/Tuning_KVM.  It should all be well known 
to the list, but a newbie is born every minute.  Please review and expand!

-- 
error compiling committee.c: too many arguments to function

Re: kvm tuning guide

[Leandro Quibem Magnabosco]

Avi Kivity escreveu:
> but a newbie is born every minute.
Reporting in!
As a newbie in KVM, I really appreciate those efforts.
Promise to help on docs when I realize that I know what I am doing. :)
Thank you.
--
Leandro Quibem Magnabosco.

Re: kvm tuning guide

[Nikola Ciprich]

"The default, IDE, is highly supported by guests but may be slow, especially with disk arrays. If your guest supports it, use the virtio interface:"
Avi,
what is the status of data integrity issues Chris Hellwig summarized some time ago?
Is it safe to recommend virtio to newbies already? Shouldn't SCSI
be safer (where applicable)?
nik


On Tue, Sep 29, 2009 at 07:30:55PM +0200, Avi Kivity wrote:
> I wrote a short tuning guide for kvm,  
> http://www.linux-kvm.org/page/Tuning_KVM.  It should all be well known  
> to the list, but a newbie is born every minute.  Please review and 
> expand!
>
> -- 
> error compiling committee.c: too many arguments to function
>
> --
> To unsubscribe from this list: send the line "unsubscribe kvm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>

-- 
-------------------------------------
Nikola CIPRICH
LinuxBox.cz, s.r.o.
28. rijna 168, 709 01 Ostrava

tel.:   +420 596 603 142
fax:    +420 596 621 273
mobil:  +420 777 093 799
www.linuxbox.cz

mobil servis: +420 737 238 656
email servis: servis@linuxbox.cz
-------------------------------------

Re: kvm tuning guide

[Avi Kivity]

On 09/30/2009 07:09 AM, Nikola Ciprich wrote:
> "The default, IDE, is highly supported by guests but may be slow, especially with disk arrays. If your guest supports it, use the virtio interface:"
> Avi,
> what is the status of data integrity issues Chris Hellwig summarized some time ago?
>    

I don't know.  Christoph?

> Is it safe to recommend virtio to newbies already?

I think so.

> Shouldn't SCSI
> be safer (where applicable)?
>    

SCSI suffers from being untested, and I think doesn't truly offer the 
parallelism it appears to.

> nik
>
>
> On Tue, Sep 29, 2009 at 07:30:55PM +0200, Avi Kivity wrote:
>    
>> I wrote a short tuning guide for kvm,
>> http://www.linux-kvm.org/page/Tuning_KVM.  It should all be well known
>> to the list, but a newbie is born every minute.  Please review and
>> expand!
>>
>> -- 
>> error compiling committee.c: too many arguments to function
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe kvm" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>      
>    


-- 
Do not meddle in the internals of kernels, for they are subtle and quick to panic.

Re: kvm tuning guide

[Christoph Hellwig]

On Wed, Sep 30, 2009 at 08:20:35AM +0200, Avi Kivity wrote:
> On 09/30/2009 07:09 AM, Nikola Ciprich wrote:
>> "The default, IDE, is highly supported by guests but may be slow, especially with disk arrays. If your guest supports it, use the virtio interface:"
>> Avi,
>> what is the status of data integrity issues Chris Hellwig summarized some time ago?
>>    
>
> I don't know.  Christoph?

On the qemu side everything is in git HEAD now, but I'm not sure about
the qemu-0.11 release as I haven't really followed it.

For the guest kernel the virtio cache flush support is now in mainline
(past-2.6.31).  For the host kernel side about 2/3 of the fixes are now
in mainline (past-2.6.31) with the others hopefully getting in this
merge window.


>> Is it safe to recommend virtio to newbies already?
>
> I think so.

I wouldn't.  At least not for people caring about their data.  It will
take a while to promote the guest side fixes to all the interesting
guests.  IDE has the major advantage that cache flush support has been
around in the guest driver for a long time so we only need to fix the
host side which is a lot easier.