----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/31051/#review72620 -----------------------------------------------------------
Ship it! Ship It! - Yusaku Sako On Feb. 14, 2015, 7:12 p.m., Gautam Borad wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/31051/ > ----------------------------------------------------------- > > (Updated Feb. 14, 2015, 7:12 p.m.) > > > Review request for Ambari, Alejandro Fernandez, Srimanth Gunturi, and Yusaku > Sako. > > > Bugs: AMBARI-9644 > https://issues.apache.org/jira/browse/AMBARI-9644 > > > Repository: ambari > > > Description > ------- > > In scenarios where Spark Job History Server is not running on same node as > ATS, the Spark service check fails during cluster bringup. > Although the JHS is up, it returns 500 to the curl requests since its waiting > for connection to ATS. > Need to add ATS also as a dependency for Spark Service Check > > > Diffs > ----- > > ambari-server/src/main/resources/stacks/HDP/2.2/role_command_order.json > 6f52cfd > > Diff: https://reviews.apache.org/r/31051/diff/ > > > Testing > ------- > > Tested on a 3 node cluster with following configuration : > > =======3======= [ NN + Clients ] > Namenode Install > ZooKeeper Server install > ZooKeeper Server start > NameNode start > > ======2======== [ S NN, HS, ATS, RM + Clients ] > ATS start *( 16:23:53 )* > History Server start ( 16:24:02 ) > RM start > Secondary NN start > HDFC Service Check > Spark Service Check *(16:25:19)* > > ======1======== [ SHS ] > Spark History Server start *( 16:23:05 )* > ZooKeeper Server start ( 16:23:10 ) > NodeManager start > > > As can be seen from the timings, Spark Service Check happens after the ATS > start. > > > Thanks, > > Gautam Borad > >
