Eric Tschetter 0f63cb4f00 1) Have IndexGeneratorJob write the descriptors for each of the segments it creates to a path in the temporary working directory (generally HDFS) ...

2) Have the DbUpdaterJob read descriptors from the temporary working directory instead of looking in the final segment output location (often the eventually consistent S3)
3) 1 and 2 Fixes #30

2012-11-20 15:30:50 -06:00

..

druid_setup.sh

1) Have IndexGeneratorJob write the descriptors for each of the segments it creates to a path in the temporary working directory (generally HDFS)

2012-11-20 15:30:50 -06:00

log4j.xml

added DruidSetup.java and install/druid_setup.sh to run it for Druid ensemble setup (put properties to zk, create zk zpaths, prep db, and dump properties/paths in zk); property druid.zk.paths.base can establish a namespace for a druid ensemble allowing more than one ensemble to use the same zookeepers; this namespace is the base zpath prefix for paths with properties in the form druid.zk.paths.*Path which are normally set automatically now; if druid.zk.paths.*Path are explicitly set, then all of them must be set with absolute paths to ensure careful attention; ZkSetup is now deprecated, use DruidSetup instead; examples/twitter RealtimeStandaloneMain will exit properly now after a kill -15 (control-c or kill pid) instead of hanging on an unstopped daemon.

2012-11-06 00:16:27 -08:00