Measuring migration downtime

Measuring migration downtime

[Arun Babu]

[-- Attachment #1.1: Type: text/plain, Size: 434 bytes --]

Hi,
 I need to get the exact downtime for doing a live migration with ms
precision.
 (I am not bothered about whole migration time which is close to 1 minute..
I need to measure
 the live migration downtime during which the DomU is out of service.. the
figure quoted in the xen
 documentation says it as ~60 ms..  )
 I hope someone must have tried to measure this before.

 Can you suggest me ideas to measure this?

Thank you,
 Arun

[-- Attachment #1.2: Type: text/html, Size: 546 bytes --]

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: Measuring migration downtime

[Anthony Liguori]

Arun Babu wrote:
> Hi,
>  I need to get the exact downtime for doing a live migration with ms 
> precision.
>  (I am not bothered about whole migration time which is close to 1 
> minute.. I need to measure
>  the live migration downtime during which the DomU is out of service.. 
> the figure quoted in the xen
>  documentation says it as ~60 ms..  )
>  I hope someone must have tried to measure this before.

If you run a benchmark against the server, and measure the server's 
throughput at a high granularity and graph it, it should show you the 
interval that it was down for.  I think that is the most realistic way 
to do this.

Regards,

Anthony Liguori

>  Can you suggest me ideas to measure this?
>
> Thank you,
> Arun
> ------------------------------------------------------------------------
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
>   

Re: Measuring migration downtime

[harish]

[-- Attachment #1.1: Type: text/plain, Size: 1324 bytes --]

Could you pls suggest any benchmarks that might have been used in the past
to do exactly this?
thanks
hmv

On 6/8/06, Anthony Liguori <aliguori@us.ibm.com> wrote:
>
> Arun Babu wrote:
> > Hi,
> >  I need to get the exact downtime for doing a live migration with ms
> > precision.
> >  (I am not bothered about whole migration time which is close to 1
> > minute.. I need to measure
> >  the live migration downtime during which the DomU is out of service..
> > the figure quoted in the xen
> >  documentation says it as ~60 ms..  )
> >  I hope someone must have tried to measure this before.
>
> If you run a benchmark against the server, and measure the server's
> throughput at a high granularity and graph it, it should show you the
> interval that it was down for.  I think that is the most realistic way
> to do this.
>
> Regards,
>
> Anthony Liguori
>
> >  Can you suggest me ideas to measure this?
> >
> > Thank you,
> > Arun
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xensource.com
> > http://lists.xensource.com/xen-devel
> >
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel
>

[-- Attachment #1.2: Type: text/html, Size: 1995 bytes --]

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: Measuring migration downtime

[Jacob Gorm Hansen]

On 6/9/06, Arun Babu <arunbabu.n@gmail.com> wrote:
>
> Hi,
>  I need to get the exact downtime for doing a live migration with ms
> precision.
>  (I am not bothered about whole migration time which is close to 1 minute..
> I need to measure
>  the live migration downtime during which the DomU is out of service.. the
> figure quoted in the xen
>  documentation says it as ~60 ms..  )
>  I hope someone must have tried to measure this before.

What about the methods and results in our NSDI 2005 paper, don't they
tell you what you need?

You can use in tcpdump on an independent machine and log timestamps of
packets send from the migrating domain's IP, in response to fx a
flood-ping.

You can also sync the two machine's clocks using NTP and instrument
the tools to measure timestamp of domain suspend resp. resume.

Jacob

Re: Measuring migration downtime

[Arun Babu]

[-- Attachment #1.1: Type: text/plain, Size: 2452 bytes --]

Jacob,
 Thanks for replying.
 I did read your paper, which gives the results in detail .
 Infact I need to do some tests exactly like the ones in your paper..(but
not in such detail)
 The method discussed estimates downtime with the idea of writable working
set & information on
 number of pages dirtied and hence evaluating the count of dirty pages to be
transfered
 during the final stop-copy phase.
 I am quite new to xen and am not sure how much effort need to be put in to
 get information on shadow page table, dirty page count etc to repeat the
tests
 you have performed ... Please let me know if we have more detailed
information
 on the method of measuring this!

 I am currently trying to estimate is based on a simple socket application
running on
 the guest domain, which does an infinite connect()-close() to a service
running on original host. When it migrates
 to the destination, it wont be able to access the service.(i hav my n/w
that way) I am using the time of last successful
 connect() call and first error in connect() to estimate
downtime.(approximately)
 This gives finer values than the ones you suggested - tcpdump with
floodping
  But I have a problem automating in my method: If migration happens exactly
after connect() call and
  gettimeofday() of a successful/failed connect, then no longer the last
succesful connect/first failed connect
  logic holds. (But always i can manually calculate with the dumps of
timestamps)

 It would be of great help if you can suggest a better method!

Sorry for a long mail!
Arun



On 6/11/06, Jacob Gorm Hansen <jacobg@diku.dk> wrote:
>
> What about the methods and results in our NSDI 2005 paper, don't they
> tell you what you need?
>
> You can use in tcpdump on an independent machine and log timestamps of
> packets send from the migrating domain's IP, in response to fx a
> flood-ping.
>
> You can also sync the two machine's clocks using NTP and instrument
> the tools to measure timestamp of domain suspend resp. resume.
>
> Jacob



On 6/9/06, Arun Babu <arunbabu.n@gmail.com> wrote:
>
> Hi,
>  I need to get the exact downtime for doing a live migration with ms
> precision.
>  (I am not bothered about whole migration time which is close to 1
minute..
> I need to measure
>  the live migration downtime during which the DomU is out of service.. the
> figure quoted in the xen
>  documentation says it as ~60 ms..  )
>  I hope someone must have tried to measure this before.

[-- Attachment #1.2: Type: text/html, Size: 3142 bytes --]

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel