webhdfs - A client library implementation for Hadoop WebHDFS, and HttpFs, for Ruby

gem install webhdfs

require 'webhdfs'
client = WebHDFS::Client.new(hostname, port)
# or with pseudo username authentication
client = WebHDFS::Client.new(hostname, port, username)

client.create('/path/to/file', data)
client.create('/path/to/file', data, :overwrite => false, :blocksize => 268435456, :replication => 5, :permission => '0666')

#This does not require whole data in memory, and it can be read chunk by chunk, ex: File data
client.create('/path/to/file', file_IO_handle, :overwrite => false, :permission => 0666)

client.append('/path/to/existing/file', data)

client.read('/path/to/target') #=> data
client.read('/path/to/target' :offset => 2048, :length => 1024) #=> data

client.mkdir('/hdfs/dirname')
client.mkdir('/hdfs/dirname', :permission => '0777')

client.rename(original_path, dst_path)

client.delete(path)
client.delete(dir_path, :recursive => true)

client.stat(file_path) #=> key-value pairs for file status
client.list(dir_path)  #=> list of key-value pairs for files in dir_path

#### To retry for LeaseExpiredException automatically
client.retry_known_errors = true

# client.retry_interval = 1 # [sec], default: 1
# client.retry_times = 1 # [times], default: 1

require 'webhdfs/fileutils'
WebHDFS::FileUtils.set_server(host, port)
# or
WebHDFS::FileUtils.set_server(host, port, username, doas)

WebHDFS::FileUtils.copy_from_local(localpath, hdfspath)
WebHDFS::FileUtils.copy_to_local(hdfspath, localpath)

WebHDFS::FileUtils.append(path, data)

client = WebHDFS::Client.new('hostname', 14000)
client.httpfs_mode = true

client.read(path) #=> data

# or with webhdfs/filetuils
WebHDFS::FileUtils.set_server('hostname', 14000)
WebHDFS::FileUtils.set_httpfs_mode
WebHDFS::FileUtils.copy_to_local(remote_path, local_path)

client = WebHDFS::Client.new('hostname', 14000, 'proxy.server.local', 8080)
client.proxy_user = 'jack'   # if needed
client.proxy_pass = 'secret' # if needed

client = WebHDFS::Client.new('hostname', 4443)
client.ssl = true
client.ssl_ca_file = "/path/to/ca_file.pem" # if needed
client.ssl_varify_mode = :peer # if needed (:none or :peer)
client.ssl_version = :TLSv1 # if needed

client = WebHDFS::Client.new('hostname', 14000)
# or if you want to use client delegation token with renewing per 8 hours and delegation token max-lifetime is 7 days
client = WebHDFS::Client.new('hostname', 14000, username, nil, nil, nil, {}, 8, 168)

client.kerberos = true
client.kerberos_keytab = "/path/to/project.keytab"

require 'openssl'

client = WebHDFS::Client.new(host, port)
client.ssl = true
client.ssl_key = OpenSSL::PKey::RSA.new(open('/path/to/key.pem'))
client.ssl_cert = OpenSSL::X509::Certificate.new(open('/path/to/cert.pem'))