> On Jan. 10, 2014, 5:04 p.m., Sean Busbey wrote: > > src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java, > > line 80 > > <https://reviews.apache.org/r/16768/diff/1/?file=419839#file419839line80> > > > > the timeout factor should get passed to the actual test runner as well > > via the -f option.
This issue will be addressed in review request# 16908 (https://reviews.apache.org/r/16908/) - Hung ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/16768/#review31517 ----------------------------------------------------------- On Jan. 9, 2014, 10:21 p.m., Hung Pham wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/16768/ > ----------------------------------------------------------- > > (Updated Jan. 9, 2014, 10:21 p.m.) > > > Review request for accumulo. > > > Bugs: ACCUMULO-2005 > https://issues.apache.org/jira/browse/ACCUMULO-2005 > > > Repository: accumulo > > > Description > ------- > > ACCUMULO-2005: Add a 3rd argument to functional test's RunTests.java to take > in as input a scaling factor for 'mapred.task.timeout'. This will allow > functional tests running as MR job to complete with the right value for > 'mapred.task.timeout' instead of failing. This scaling factor, called > 'timeout_factor', is an integer indicating multiples of the default value of > 480 secs for 'mapred.task.timeout'. For instances, a timeout_factor of 1 > means 480 secs, and 2 would be 960 secs, and so on. > > Testing was done on small cluster (4 nodes: master + 3 workers) to ensure MR > job complete successfully instead of failing out due to timeouts that lead to > killed attempts. > > > Diffs > ----- > > > src/server/src/main/java/org/apache/accumulo/server/test/functional/RunTests.java > 1a19bb5 > test/system/auto/README 4773393 > > Diff: https://reviews.apache.org/r/16768/diff/ > > > Testing > ------- > > Testing was done on small cluster (4 nodes: master + 3 workers) to ensure MR > job complete successfully instead of failing out due to timeouts that lead to > killed attempts. > > > Thanks, > > Hung Pham > >
