The Master doesn't do work. I don't quite understand the rest; there's not a "Spark slave" role. You can have master and workers, and even your driver, on one machine; what's the error?
On Sun, Aug 31, 2014 at 12:53 AM, arthur.hk.c...@gmail.com <arthur.hk.c...@gmail.com> wrote: > Hi, > > I have few questions about Spark Master and Slave setup: > > Here, I have 5 Hadoop nodes (n1, n2, n3, n4, and n5 respectively), at the > moment I run Spark under these nodes: > n1: Hadoop Active Name node, Hadoop Slave Spark Active Master > n2: Hadoop Standby Name Node, Hadoop Salve Spark Slave > n3: Hadoop Salve Spark Slave > n4: Hadoop Salve Spark Slave > n5: Hadoop Salve Spark Slave > > Questions: > Q1: If I set n1 as both Spark Master and Spark Slave, I cannot start the > Spark Cluster. does it mean that, unlike Hadoop, I cannot use the same > machine to be both MASTER and SLAVE in Spark? > n1: Hadoop Active Name node, Hadoop Slave Spark Active Master Spark > Slave (failed to Start Spark) > n2: Hadoop Standby Name Node, Hadoop Salve Spark Slave > n3: Hadoop Salve Spark Slave > n4: Hadoop Salve Spark Slave > n5: Hadoop Salve Spark Slave > > Q2: I am planning Spark HA, what if I use n2 as "Spark Standby Master and > Spark Slave”? is Spark allowed to run Standby Master and Slave under same > machine? > n1: Hadoop Active Name node, Hadoop Slave Spark Active Master > n2: Hadoop Standby Name Node, Hadoop Salve Spark Standby Master Spark > Slave > n3: Hadoop Salve Spark Slave > n4: Hadoop Salve Spark Slave > n5: Hadoop Salve Spark Slave > > Q3: Does the Spark Master node do actual computation work like a worker or > just a pure monitoring node? > > Regards > Arthur --------------------------------------------------------------------- To unsubscribe, e-mail: user-unsubscr...@spark.apache.org For additional commands, e-mail: user-h...@spark.apache.org
