----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/29442/#review66199 -----------------------------------------------------------
Ship it! This looks good. - Mahadev Konar On Dec. 27, 2014, 8:13 p.m., Robert Nettleton wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/29442/ > ----------------------------------------------------------- > > (Updated Dec. 27, 2014, 8:13 p.m.) > > > Review request for Ambari, John Speidel, Mahadev Konar, and Nate Cole. > > > Repository: ambari > > > Description > ------- > > This patch fixes AMBARI-8934. > > Recent changes to the Hive stack definitions in Ambari have modified the > ordering in which services start up in Hive. > > In particular, the MYSQL_SERVER component appears to startup during the > INSTALL phase of a cluster deployment. > > This causes the ambari-server HeartbeatHandler to detect that the > MYSQL_SERVER is up and running earlier, and the handler > will then set the state directly on the ServiceComponentHost. Setting the > state directly causes errors in the event handling in Ambari, and a > ServiceComponentHostEvent request is handled to start the MYSQL_SERVER after > it has already been started. > > This patch addresses this issue by modifing the ServiceComponentHost event > handling code, such that requests for a component "START" are not considered > failures if the component has already been started. > > > Diffs > ----- > > > ambari-server/src/main/java/org/apache/ambari/server/state/cluster/ClusterImpl.java > f80044e > > Diff: https://reviews.apache.org/r/29442/diff/ > > > Testing > ------- > > 1. Manually verified that a single-node HDP cluster that includes Hive can be > deployed via a Blueprint. > 2. Ran the ambari-server unit tests: > "Results : > > Tests run: 2468, Failures: 0, Errors: 0, Skipped: 13" > > > Thanks, > > Robert Nettleton > >
