----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/30986/#review72361 -----------------------------------------------------------
Ship it! Ship It! - Vitalyi Brodetskyi On Лют. 13, 2015, 11:20 до полудня, Andrew Onischuk wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/30986/ > ----------------------------------------------------------- > > (Updated Лют. 13, 2015, 11:20 до полудня) > > > Review request for Ambari and Vitalyi Brodetskyi. > > > Bugs: AMBARI-9618 > https://issues.apache.org/jira/browse/AMBARI-9618 > > > Repository: ambari > > > Description > ------- > > Add checks in Ambaripreupload.py so that there is minimum chances of tarball > uploads in parallel. > > We should add the following checks in: > > > > > > if dir exists wasb:///hdp/apps/{{ hdp_stack_version }}/mapreduce/: > copy_tarballs_to_hdfs("/usr/hdp/current/hadoop-client/mapreduce.tar.gz", > "wasb:///hdp/apps/{{ hdp_stack_version }}/mapreduce/", > 'hadoop-mapreduce-historyserver', params.mapred_user, params.hdfs_user, > params.user_group) > > copy_tarballs_to_hdfs("/usr/hdp/current/sqoop-client/sqoop.tar.gz", > "wasb:///hdp/apps/{{ hdp_stack_version }}/sqoop/", > 'hadoop-mapreduce-historyserver', params.mapred_user, params.hdfs_user, > params.user_group) > > > > > We should do this with all copy tarballs to hdfs in the file - meaning checks > if the directories exist and then only copy the tarballs. > > The reason for doing this is so that when Ambari is doing the install/upload > tarballs there is less chances of the ambaripreupload.py script creating a > race condition on tarball upload. > > > Diffs > ----- > > ambari-server/src/main/resources/scripts/Ambaripreupload.py 89e0742 > > Diff: https://reviews.apache.org/r/30986/diff/ > > > Testing > ------- > > mvn clean test > > > Thanks, > > Andrew Onischuk > >
