Re: [PATCH RFC 3/3] qtest/migration: Use switchover-hold to speedup

[Peter Xu]

On Fri, Jun 02, 2023 at 12:03:05PM +0100, Daniel P. Berrangé wrote:
> On Thu, Jun 01, 2023 at 09:16:26PM -0400, Peter Xu wrote:
> > Use the switchover-hold flag rather than tuning bw+downtime to guide test
> > convergence.
> > 
> > This can achieve similar goal of previous patch "tests/qtest: massively
> > speed up migration-test" but without magic offset to write or monitoring.
> > 
> > The initial solution can reduce migration-test time from 8min to 1min40s,
> > this patch can further reduce it from 1m40s to 1m1s per my local test.
> > 
> > Signed-off-by: Peter Xu <peterx@redhat.com>
> > ---
> >  tests/qtest/migration-test.c | 20 ++++++++++++++------
> >  1 file changed, 14 insertions(+), 6 deletions(-)
> > 
> > diff --git a/tests/qtest/migration-test.c b/tests/qtest/migration-test.c
> > index b0c355bbd9..62bdd67fd9 100644
> > --- a/tests/qtest/migration-test.c
> > +++ b/tests/qtest/migration-test.c
> > @@ -433,16 +433,15 @@ static void migrate_set_parameter_bool(QTestState 
> > *who, const char *parameter,
> >  
> >  static void migrate_ensure_non_converge(QTestState *who)
> >  {
> > -    /* Can't converge with 1ms downtime + 3 mbs bandwidth limit */
> > -    migrate_set_parameter_int(who, "max-bandwidth", 3 * 1000 * 1000);
> > -    migrate_set_parameter_int(who, "downtime-limit", 1);
> > +    /* Hold off switchover for precopy only */
> > +    migrate_set_parameter_bool(who, "switchover-hold", true);
> >  }
> >  
> >  static void migrate_ensure_converge(QTestState *who)
> >  {
> > -    /* Should converge with 30s downtime + 1 gbs bandwidth limit */
> > -    migrate_set_parameter_int(who, "max-bandwidth", 1 * 1000 * 1000 * 
> > 1000);
> > -    migrate_set_parameter_int(who, "downtime-limit", 30 * 1000);
> > +    /* No limitation on bandwidth so converge faster */
> > +    migrate_set_parameter_int(who, "max-bandwidth", 0);
> 
> You're already setting max-bandwith==0 in migrate_postcopy_prepare.
> 
> If you also set it in test_precopy_common, so we don't need to
> set it here, and we'll avoid the initial phase running with
> bandwidth=32mbs.

There are more tests than test_precopy_common() that leverages this, so I
used to kept it there to still share some code:

test_migrate_auto_converge[2021] migrate_ensure_converge(from);
test_multifd_tcp_cancel[2349]  migrate_ensure_converge(from);

But I think you're right - moving the bw=0 setup out should be cleaner.
I'll see whether I should just move it into all the specific tests.

Thanks,

-- 
Peter Xu