On Sun, Oct 17, 2010 at 9:05 PM, Ted Dunning <ted.dunn...@gmail.com> wrote: > Over the last two days Mahout-Quality has had 4 successful builds and no > failures from solaris1 and 7 builds with 4 failures from solaris2. A quick > chi^2 test indicates that this is a significant difference at p < 0.02. > > I think solaris2 has something wrong. I also think that this changed two > days ago.
>From the logs, this seems to be related to the Clover problem reported a few days ago. Could you try to disable Clover to see if that helps? At least that would help us isolate the problem. Also, I've restarted the solaris2 slave. /niklas
