On Fri, Aug 31, 2012 at 8:48 AM, Ted Yu <yuzhih...@gmail.com> wrote: > Thanks for the response Dave. > > I exchanged a few emails with Lars H last night and we felt we shouldn't > block 0.92.2 / 0.94.2 releases due to a few flaky tests. > > Let's fix flaky tests continuously. >
That seems basically fine except I'd think that we'd pass most of the time rather than as it is now, where we fail and its the exception that tests pass. Whatever about Jenkins, I'd think that when I build locally, again, most of the time the build should pass but its not been the case during my builds of 0.92 locally. Has triage on 0.92 failing tests been done to figure if tests are flakey or whether its hbase that is the problem (I've seen JIRA traffic on 0.92 unit tests so this may have been done already). Can we not disable tests that we've labelled flakey and not issues in hbase? St.Ack
