Re: [PATCH v3 8/9] tests/qtest: make more migration pre-copy scenarios run non-live

[Daniel P . Berrangé]

On Thu, Jun 01, 2023 at 11:30:10AM -0400, Peter Xu wrote:
> Thanks for looking into this.. definitely worthwhile.
> 
> On Wed, May 31, 2023 at 02:23:59PM +0100, Daniel P. Berrangé wrote:
> > There are 27 pre-copy live migration scenarios being tested. In all of
> > these we force non-convergance and run for one iteration, then let it
> > converge and wait for completion during the second (or following)
> > iterations. At 3 mbps bandwidth limit the first iteration takes a very
> > long time (~30 seconds).
> > 
> > While it is important to test the migration passes and convergance
> > logic, it is overkill to do this for all 27 pre-copy scenarios. The
> > TLS migration scenarios in particular are merely exercising different
> > code paths during connection establishment.
> > 
> > To optimize time taken, switch most of the test scenarios to run
> > non-live (ie guest CPUs paused) with no bandwidth limits. This gives
> > a massive speed up for most of the test scenarios.
> > 
> > For test coverage the following scenarios are unchanged
> 
> Curious how are below chosen?  I assume..

Chosen based on whether they exercise code paths that are unique
and interesting during the RAM transfer phase.

Essentially the goal is that if we have N% code coverage before this
patch, then we should still have the same N% code coverage after this
patch.

The TLS tests exercise code paths that are unique during the migration
establishment phase. Once establishd they don't exercise anything
"interesting" during RAM transfer phase. Thus we don't loose code coverage
by runing TLS tests non-live.

> 
> > 
> >  * Precopy with UNIX sockets
> 
> this one verifies dirty log.
> 
> >  * Precopy with UNIX sockets and dirty ring tracking
> 
> ... dirty ring...
> 
> >  * Precopy with XBZRLE
> 
> ... xbzrle I think needs a diff on old/new, makes sense.
> 
> >  * Precopy with UNIX compress
> >  * Precopy with UNIX compress (nowait)
> >  * Precopy with multifd
> 
> What about the rest three?  Especially for two compression tests.

The compress thread logic is unique/interesting during RAM transfer
so benefits from running live. The wait vs non-wait scenario tests
a distinct codepath/logic.

multifd has been a massive source of bugs and has very different
codepaths from non-multifd, so is needed for codeage.


With regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|