Hi, If I have TWO nodes for ResourceManager HA, what should be the correct steps and commands to start and stop ResourceManager in a ResourceManager HA cluster ? Unlike ./sbin/start-dfs.sh (which can start all NNs from a NN), it seems that ./sbin/start-yarn.sh can only start YARN in a node at a time.
Regards Arthur On 11 Aug, 2014, at 11:04 pm, [email protected] <[email protected]> wrote: > Hi > > I am running Hadoop 2.4.1 with YARN HA enabled (two name nodes, NM1 and NM2). > When verifying ResourceManager failover, I use “kill -9” to terminate the > ResourceManager in name node 1 (NM1), if I run the the test job, it seems > that the failover of ResourceManager keeps trying NM1 and NM2 non-stop. > > Does anyone have the idea what would be wrong about this? Thanks > > Regards > Arthur > > > > bin/hadoop jar share/hadoop/mapreduce/hadoop-mapreduce-examples-2.4.1.jar pi > 5 1010000000 > Number of Maps = 5 > Samples per Map = 1010000000 > Wrote input for Map #0 > Wrote input for Map #1 > Wrote input for Map #2 > Wrote input for Map #3 > Wrote input for Map #4 > Starting Job > 14/08/11 22:35:23 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm2 > 14/08/11 22:35:24 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm1 > 14/08/11 22:35:25 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm2 > 14/08/11 22:35:28 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm1 > 14/08/11 22:35:30 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm2 > 14/08/11 22:35:32 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm1 > 14/08/11 22:35:34 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm2 > 14/08/11 22:35:37 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm1 > 14/08/11 22:35:39 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm2 > 14/08/11 22:35:40 INFO client.ConfiguredRMFailoverProxyProvider: Failing over > to nm1 > ….
