In the zoo.cfg the list of servers you provide are the zookeeper quorum. Zookeeper is a distributed service that usually runs on 3 or 5 nodes in large/production settings.
Because your zookeeper is expecting a quorum of duff servers, but only ever gets one, it never achieves a "majority." So the tablet servers can't obtain locks because zookeeper is not giving any out, and thus don't start. Replace your server lines with just server.1=zookeeper1:2181 and that should work better. On Apr 29, 2014 6:07 AM, "Marko Escriba" <[email protected]> wrote: > Hi, > > I am configuring Accumulo Instance in a multinode cluster (AWS Environment) > with Hadoop Namenode, Secondary Namenode, Zookeeper Server and 3 > Datanode/Slaves running on each separate instances. I have set Namenode and > Secondary Namenode as Accumulo Masters and the 3 Datanodes as Slaves. > > In zoo.cfg (Zookeeper instance): > > clientPort=2181 > Server.1=Master1:2181 > Server.2=Master2:2181 > Server.3=Node01:2181 > Server.4=Node02:2181 > Server.5=Node03:2181 > > and in accumulo-site.xml: > > <property> > <name>instance.zookeeper.host</name> > <value>Zookeeper:2181</value> > <description>comma separated list of zookeeper servers</description> > </property> > > the Zookeeper:2181 is the host (instance) which the zookeeper service is > running. > > I would like to ask for guidance if I am doing it right? This is my first > time configuring accumulo in multinode. > And I am lost. When I initialize accumulo, it create an instance and lock > the master to "Master1" node, but the tservers are down (3 slaves). > > Any help please. Thanks in advance. > > > > > > > -- > View this message in context: > http://apache-accumulo.1065345.n5.nabble.com/Accumulo-tservers-tp9485.html > Sent from the Developers mailing list archive at Nabble.com. >
