> On Feb. 6, 2015, 6:43 p.m., Jayush Luniya wrote: > > ambari-common/src/main/python/resource_management/libraries/functions/install_jdbc_driver.py, > > line 42 > > <https://reviews.apache.org/r/30702/diff/1/?file=851754#file851754line42> > > > > We should just copy the JDBC driver from the cache path. This ensures > > same version of the driver is copied over in the classpath.
1. Leaving the driver copied only from the cache would be inconsistent with the JDBC connection test, which downloads the driver from the resources dir. If the user skipped the JDBC connection test (possible, under warning) and we didn't download the driver, the installation will fail. 2. (less important, but not negligible) This approach opens the possibility to deploy the driver from a configured URL (same as the JDBC connection test), sparing the server some traffic, and facilitating a potential future upgrade. > On Feb. 6, 2015, 6:43 p.m., Jayush Luniya wrote: > > ambari-server/src/main/resources/stacks/HDPWIN/2.1/hooks/after-INSTALL/scripts/hook.py, > > line 36 > > <https://reviews.apache.org/r/30702/diff/1/?file=851755#file851755line36> > > > > I dont see a need to add another function for this > > > > See : > > ambari-server/src/main/resources/stacks/HDPWIN/2.1/services/OOZIE/package/scripts/oozie_server.py > > > > > > force_download_file(os.path.join(params.config['hostLevelParams']['jdk_location'], > > "sqljdbc4.jar"), > > os.path.join(params.oozie_root, "extra_libs", "sqljdbc4.jar") > > ) 1. Code reuse vs copy-paste. 2. Actually, it's Oozie who's out of line. It uses its own copy of the driver, while it was predictable that the driver would be a shared component, it forces a download over a previously downloaded file, and besides it doesn't even bother to download the integrated authentication dll (which will be needed when we start using integrated authentication with domain accounts). - Florian ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/30702/#review71469 ----------------------------------------------------------- On Feb. 6, 2015, 12:03 a.m., Florian Barca wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/30702/ > ----------------------------------------------------------- > > (Updated Feb. 6, 2015, 12:03 a.m.) > > > Review request for Ambari, Artem Baranchuk, Eugene Chekanskiy, Jayush Luniya, > and Mahadev Konar. > > > Bugs: AMBARI-9505 > https://issues.apache.org/jira/browse/AMBARI-9505 > > > Repository: ambari > > > Description > ------- > > Copying the SQL Server JDBC driver into the hadoop shared lib directory. > > > Diffs > ----- > > > ambari-common/src/main/python/resource_management/libraries/functions/__init__.py > 9b32b92cc8845d775c89f1715f4ae8d98c365843 > > ambari-common/src/main/python/resource_management/libraries/functions/install_jdbc_driver.py > PRE-CREATION > > ambari-server/src/main/resources/stacks/HDPWIN/2.1/hooks/after-INSTALL/scripts/hook.py > 31df28ddb174a689717978de1cb3fce8603cd11b > > Diff: https://reviews.apache.org/r/30702/diff/ > > > Testing > ------- > > mvn clean install > Multi-node deployment on a local VM cluster -> Hive deploys and starts ok > > > Thanks, > > Florian Barca > >
