[ovs-dev] [PATCH] lacp.at: Change timing of lacp - negotiation test

Ben Pfaff blp at nicira.com
Thu Mar 28 18:48:17 UTC 2013


On Thu, Mar 28, 2013 at 08:37:27AM -0700, Ben Pfaff wrote:
> On Thu, Mar 28, 2013 at 07:57:47AM -0700, Ben Pfaff wrote:
> > On Thu, Mar 28, 2013 at 03:47:46PM +0200, Jarno Rajahalme wrote:
> > > The "lacp - negotiation" test (# 3) used to always fail when running
> > > multiple tests in parallel (-j2).  The reduced simulated wait time
> > > (from 4 seconds to 2.5 seconds) seems to still give the expected result,
> > > and works also when running multiple tests in parallel.
> > > 
> > > ---
> > > Tested with upto -j4, x2 the cores in the test laptop.
> > > Maybe someone with a really fast machine needs to check that this
> > > test still works.
> > > 
> > > Signed-off-by: Jarno Rajahalme <jarno.rajahalme at nsn.com>
> > 
> > Thanks!  I applied this to master.
> > 
> > With this patch, the test passes for me, with -j8, on my fast and my
> > slow machines.  Previously it only passed on my fast machine.  (I should
> > have done something about it a long time ago, but I don't use the slow
> > machine much for dev.)
> 
> But now it fails on our internal autobuilder servers, where it always
> passed before.
> 
> I'm thinking about just marking this test to be skipped for now, until
> we can figure out why it has a timing dependency.  Any objection?

I posted a patch that I hope really fixes the timing dependency:
http://openvswitch.org/pipermail/dev/2013-March/026309.html



More information about the dev mailing list