----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/50407/#review143472 -----------------------------------------------------------
ambari-common/src/main/python/resource_management/libraries/script/script.py (lines 369 - 372) <https://reviews.apache.org/r/50407/#comment209274> Is there any reason we shouldn't just explicitely trust `commandParams/version` if it's supplied? Does `hostLevelParams/package_version` take precedence? - Jonathan Hurley On July 25, 2016, 4:26 p.m., Nate Cole wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/50407/ > ----------------------------------------------------------- > > (Updated July 25, 2016, 4:26 p.m.) > > > Review request for Ambari, Alejandro Fernandez, Dmitro Lisnichenko, and > Jonathan Hurley. > > > Bugs: AMBARI-17883 > https://issues.apache.org/jira/browse/AMBARI-17883 > > > Repository: ambari > > > Description > ------- > > STR: > 1. Install ZK only cluster with HDP-2.4 > 2. Install repo for HDP-2.5, and distribute bits > 3. Do not upgrade, but then add HDFS. > > The system will pick HDP-2.5 which will fail. Fix is to check for > roleCommand == 'INSTALL' and the passed json 'version' to be present, which > is true after initial install. > > > Diffs > ----- > > > ambari-common/src/main/python/resource_management/libraries/script/script.py > 5deeb52 > ambari-server/src/test/python/stacks/2.0.6/HBASE/test_hbase_master.py > f5ed727 > > Diff: https://reviews.apache.org/r/50407/diff/ > > > Testing > ------- > > Manual testing of above STR, and newly added test: > > ---------------------------------------------------------------------- > Total run:1033 > Total errors:0 > Total failures:0 > OK > > > Thanks, > > Nate Cole > >
