Question about namenode HA

2014-12-05 Thread Alexandru Pacurar
Hello, I'm trying to configure HA for the HDFS namenode with QJM following the instructions form here http://hadoop.apache.org/docs/r2.2.0/hadoop-yarn/hadoop-yarn-site/HDFSHighAvailabilityWithQJM.html. My setup is the following : Ubuntu 12.04.5 LTS on all the nodes, Hadoop 2.4.1 installed, two

Re: Question about namenode HA

2012-10-17 Thread Chao Shi
Get we get rid of ZK completely? Since JNs are like simplified version of ZK, it should be possible to use it for election. I think it pretty easy: - JN exposes latest heartbeat information via RPC (the active NN heart-beats JNs every 1 second) - zkfc decided whether the current active NN is alive

Re: Question about namenode HA

2012-10-15 Thread Todd Lipcon
Hi Liang, Answers inline below. On Sun, Oct 14, 2012 at 8:01 PM, 谢良 wrote: > Hi Todd and other HA experts, > > I've two question: > > 1) why the zkfc is a seperate process, i mean, what's the primary design > consideration that we didn't integrate zkfc features into namenode self ? > > There ar

Re: Question about namenode HA

2012-10-15 Thread Steve Loughran
On 15 October 2012 04:01, 谢良 wrote: > Hi Todd and other HA experts, > > I've two question: > > 1) why the zkfc is a seperate process, i mean, what's the primary design > consideration that we didn't integrate zkfc features into namenode self ? > I don't know the answer to that, except I do know

Question about namenode HA

2012-10-14 Thread 谢良
Hi Todd and other HA experts, I've two question: 1) why the zkfc is a seperate process, i mean, what's the primary design consideration that we didn't integrate zkfc features into namenode self ? 2) If i deploy CDH4.1(included QJM feature), since QJM can do fencing writer, so can i just conf