On Tue, Mar 18, 2014 at 11:42 PM, haosdent <haosd...@gmail.com> wrote:
> For hudson console: > > [INFO] HBase - Server .................................... FAILURE > [1:03:34.814s] > > [ERROR] Failed to execute goal > org.apache.maven.plugins:maven-surefire-plugin:2.12-TRUNK-HBASE-2:test > (secondPartTestsExecution) on project hbase-server: Failure or timeout > -> [Help 1] > > > It looks like TestHCM takes too long and build failed, as HBASE-10787. > > The above snippet says nought about TestHCM? Making tests take less time is a fine endeavor. I am just trying to follow how TestHCM is being fingered as the zombie. Do we have stack traces catching it in its zombie role? Usually we add timeout annotations to suspected go-zombie tests so they'll timeout rather than go zombie. I do not see that in the patch IIRC. St.Ack
