A total of four virtual machine including 2 sets the namenode, three datanodes, three zookeeper
Zookeeper cluster for the first time after the success of the building can be normal use, but the server restart zookeeper cannot successfully launched,
folder to create, zoo. The CFG profile information as well as the myid id works,
firewall has shut down,
Open the zookeeper log to see problems,
In/usr/local/SRC/jdk1.6.0 _45/bin/Java this path is under the hadoop etc before yarn - env. Sh and mapred - env. Sh the default path in front of the manual configuration;
Configuration path is as follows:
Above is the Java path before change log errors in path
midway accidentally deleted a namenode server, and then according to the steps to put this one namenode returns the zookeeper service can normal boot, but the cluster to restart again after this problem, the
The namenode and yarn in the datanode - env. Sh and mapred - env. Sh the same configuration,
To the current problem is no clue, solutions,
CodePudding user response:
Resolved, after browsing online for a long time, found that should be the zk service cannot find javahome and error, and then I will go to the zookeeper package under the bin folder zkEnv. Found in sh file to identify the path to that part of the configuration code, to the absolute path, and then distributed to several other zk server, after the restart normal service,
Commented out is the original code, restore the javahome into an absolute path,