> >  > OK I've made that change.  I'm running the stress test again just to be
 > >  > safe.  I'll also re-run the DAD test suite.
 > > 
 > > Well, I tried to re-run the DAD test suite but many of the test cases have
 > > started to fail, possibly due to some issue with psh.  I see the same
 > > failures with a stock Nevada nightly.   Most of the failures look like:
 > > 
 > >   Test_Case_Start| 102320 tests/test-v4static-1 | 02:53:01 0 |
 > >   stdout| ERROR: failed to receive expected ARP traffic
 > >   stdout| Received 0 probes (wanted 3)
 > 
 > I've seen behavior like that on some blade systems, where the local
 > embedded switch has a (hard-coded? ARP only?) broadcast limit.

This was a T1000.

 > In any event, the change you're looking at should have nothing to do
 > with a failure like that.

Right.  It's just frustrating.

-- 
meem

Reply via email to