Database Reference
In-Depth Information
describes installing the ZooKeeper server for use in either a standalone
development environment or in a multiserver cluster.
Installing the ZooKeeper Server
ZooKeeper can often be found in the installation packages of Hadoop
distributions. These distributions often have repositories compatible with
various server flavors.
It is also possible to install ZooKeeper from the binary archives available
on the Apache ZooKeeper website (
http://zookeeper.apache.org
)
. The most
current version of the server at the time of writing is version 3.4.5. Generally
speaking, this latest version works with most applications, but it introduces
some differences relative to the 3.3 series of servers. For software that
requires the 3.3 series of ZooKeeper, 3.3.6 is the latest stable version.
After unpacking the ZooKeeper tarball, the directory structure should look
something like this:
$ cd zookeeper-3.4.5/
$ ls
CHANGES.txt docs
LICENSE.txt ivy.xml
NOTICE.txt ivysettings.xml
README.txt lib
README_packaging.txt recipes
Bin src
build.xml zookeeper-3.4.5.jar
conf zookeeper-3.4.5.jar.asc
contrib zookeeper-3.4.5.jar.md5
dist-maven zookeeper-3.4.5.jar.sha1
Before the server can be started, it must be configured. There is a sample
configuration in the
conf
directory that provides a good starting point
for configuring a ZooKeeper server. Start by copying that file to
zoo.cfg
,
which is what the server startup script expects the configuration file to be
called. The first few options in this file usually do not require modification,
as they have to do with the time between heartbeat packets:
# The number of milliseconds of each tick
tickTime=2000
