Hitachi Hyper Scale-Out Platform (HSP) MK-95HSP017-03 11 October 2016
2016 Hitachi, Ltd. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying and recording, or stored in a database or retrieval system for commercial purposes without the express written permission of Hitachi, Ltd., or Hitachi Data Systems Corporation (collectively, Hitachi ). Licensee may make copies of the Materials provided that any such copy is: (i) created as an essential step in utilization of the Software as licensed and is used in no other manner; or (ii) used for archival purposes. Licensee may not make any other copies of the Materials. "Materials" mean text, data, photographs, graphics, audio, video and documents. Hitachi reserves the right to make changes to this Material at any time without notice and assumes no responsibility for its use. The Materials contain the most current information available at the time of publication. Some of the features described in the Materials might not be currently available. Refer to the most recent product announcement for information about feature and product availability, or contact Hitachi Data Systems Corporation at https://support.hds.com/en_us/contact-us.html. Notice: Hitachi products and services can be ordered only under the terms and conditions of the applicable Hitachi agreements. The use of Hitachi products is governed by the terms of your agreements with Hitachi Data Systems Corporation. By using this software, you agree that you are responsible for: 1) Acquiring the relevant consents as may be required under local privacy laws or otherwise from authorized employees and other individuals to access relevant data; and 2) Verifying that data continues to be held, retrieved, deleted, or otherwise processed in accordance with relevant laws. Notice on Export Controls. The technical data and technology inherent in this Document may be subject to U.S. export control laws, including the U.S. Export Administration Act and its associated regulations, and may be subject to export or import regulations in other countries. Reader agrees to comply strictly with all such regulations and acknowledges that Reader has the responsibility to obtain licenses to export, re-export, or import the Document and any Compliant Products. Hitachi is a registered trademark of Hitachi, Ltd., in the United States and other countries. AIX, AS/400e, DB2, Domino, DS6000, DS8000, Enterprise Storage Server, eserver, FICON, FlashCopy, IBM, Lotus, MVS, OS/390, PowerPC, RS6000, S/390, System z9, System z10, Tivoli, z/os, z9, z10, z13, z/vm, and z/vse are registered trademarks or trademarks of International Business Machines Corporation. Active Directory, ActiveX, Bing, Excel, Hyper-V, Internet Explorer, the Internet Explorer logo, Microsoft, the Microsoft Corporate Logo, MS-DOS, Outlook, PowerPoint, SharePoint, Silverlight, SmartScreen, SQL Server, Visual Basic, Visual C++, Visual Studio, Windows, the Windows logo, Windows Azure, Windows PowerShell, Windows Server, the Windows start button, and Windows Vista are registered trademarks or trademarks of Microsoft Corporation. Microsoft product screen shots are reprinted with permission from Microsoft Corporation. All other trademarks, service marks, and company names in this document or web site are properties of their respective owners.
Contents Preface Intended audience Product version Release notes Document revision level Document conventions Accessing product documentation Getting help Comments v vi vi vi vi vii vii viii viii Deploy the Ambari VMs on HSP 1 Before you begin 2 2 Change the ssh key (optional steps) 4 Deploy the Ambari template 6 Install script package 6 Setup Ambari 7 8 Configure the new cluster 9 10 Contents iii
iv Contents
Preface This document describes and provides instructions for deploying the Ambari 2.2.2 and HDP 2.4 VM template on Hitachi Hyper Scale-Out Platform (HSP). This Preface includes the following information: Intended audience Product version Document revision level Document conventions Accessing product documentation Getting help Comments Preface v
Intended audience Intended audience This document is intended for system administrators, Hitachi Data Systems representatives, and authorized service providers who need to deploy the Ambari 2.2.2 and HDP 2.4 VM template on Hitachi Hyper Scale-Out Platform (HSP). Readers of this document should be familiar with the following: Ambari administration Linux operating system and working in a restricted shell environment Site-specific network information Product version This document applies to Hyper Scale-Out Platform release 1.2.0.037 or later. Release notes The release notes for this product are available on Hitachi Data Systems Support Connect: https://support.hds.com/en_us/contact-us.html. Read the release notes before installing and using this product. They may contain requirements or restrictions that are not fully described in this document or updates or corrections to this document. Document revision level Revision Date Description MK-95HSP017-00 April 2016 Initial release MK-95HSP017-01 May 2016 01, supersedes and replaces MK-95HSP017-00 MK-95HSP017-02 June 2016 02, supersedes and replaces MK-95HSP017-01 MK-95HSP017-03 11 October 2016 03, supersedes and replaces MK-95HSP017-02 vi Preface
Document conventions Document conventions This document uses the following typographic conventions: Convention Description Bold Indicates: Text in a window or dialog box, such as menus, menu options, buttons, and labels. For example: On the Add Pair dialog box, click OK. Text appearing on screen or entered by the user. For example: The -split option <Italic> in angle brackets monospace Indicates a variable, which is a placeholder for site- or installationspecific details that you need to provide. For example: copy <source-file> <target-file> Indicates text that is displayed on the screen or text that you need to enter. For example: # pairdisplay -g oradb Also, the name of a directory, folder, or file. For example: The horcm.conf file This document uses the following icons to draw attention to information: Icon Meaning Description Important Calls attention to information that is essential to the completion of a task. Caution Warns that failure to take or avoid a specified action could result in adverse conditions or consequences (for example, loss of access to data). Accessing product documentation Product documentation is available on Hitachi Data Systems Support Connect: https://knowledge.hds.com/documents. Check this site for the most current documentation, including important updates that may have been made after the release of the product. Preface vii
Getting help Getting help Hitachi Data Systems Support Portal is the destination for technical support of products and solutions sold by Hitachi Data Systems. To contact technical support, log on to Hitachi Data Systems Support Connect for contact information: https://support.hds.com/en_us/contact-us.html. Hitachi Data Systems Community is a global online community for HDS customers, partners, independent software vendors, employees, and prospects. It is the destination to get answers, discover insights, and make connections. Join the conversation today! Go to community.hds.com, register, and complete your profile. Comments Please send us your comments on this document: hsp.documentation.comments@hds.com Include the document title and part number, including the revision (for example, -01), and refer to specific sections and paragraphs whenever possible. All comments become the property of Hitachi Data Systems. Thank you! viii Preface
1 Deploy the Ambari VMs on HSP This chapter describes the procedure for deploying the Ambari 2.2.2 cluster. At the end of this procedure the VM instances will be deployed and ready to be configured. Before you begin Deploy the Ambari VMs on HSP 1
Before you begin Before you begin Verify that you have: An initialized HSP cluster running at least HSP 1.2.0.037. Ensure that your HSP cluster has its DNS domain name, DNS server and NTP server configured correctly. Downloaded copy of the tgz file containing the vm-template for Ambari from Hortonworks. This file is available from Technical Information Service Center (TISC), and may be called something like CentOS-7.2_ HSP-1.2_r1.tgz Downloaded copy of the TGZ file containing the repository-vm for Ambari from Hortonworks. This file is available from Technical Information Service Center (TISC), and may be called something like Repo-with_HDP-2.4_Ambari-2.2.2_r1.tgz Linux client. First, add the vm-template to the HSP cluster. 1. Extract the two tgz files so that you have an ISO file for the vm-template and an ISO file for the repository VM template. 2. Extract the qcow2 images from the ISO files. On the Linux client, mount the ISO image as shown below: linux# mount t iso9660 o loop /tmp/centos-7.2_hsp-1.2_r1.iso /mnt/iso linux# mount t iso9660 o loop /tmp/repo-with_hdp-2.4_ambari-2.2.2_ r1.iso /mnt/iso 3. Mount a share from the HSP cluster on your Linux system: linux# mount <HSP Cluster Virtual IP Address>:/<myshare> /mnt/hsp Where /myshare is the name of the available share. 4. Copy the qcow2 images from the ISOs to the share: linux# cp /mnt/iso/centos-7.2_hsp-1.2_r1.qcow2 /mnt/hsp linux# cp /mnt/iso/repo-with_hdp-2.4_ambari-2.2.2_r1.qcow2 /mnt/hsp 2 Deploy the Ambari VMs on HSP
If you will be powering on the template in sandbox mode (for example, to Change the ssh key (optional steps)) the CentOS qcow2 image needs to be writeable: linux# chmod 666 /mnt/hsp/centos-7.2_hsp-1.2_r1.qcow2 5. On the HSP cluster, add the vm-template and RepositoryVM-template: admin@node003:~$ hspadm vm-template add --name ambari_template1 -- image-format qcow2 --image-path <myshare>:/centos-7.2_hsp-1.2_ r1.qcow2 --use-address-pool Y admin@node003:~$ hspadm vm-template add --name RepositoryVM_template --image-format qcow2 --image-path <myshare>:/repo-with_hdp-2.4_ Ambari-2.2.2_r1.qcow2 6. Verify this was successful: admin@node003:~$ hspadm vm-template list 7. Deploy a single instance of the Repository-VM: admin@node003:~$ hspadm vm-instance add --name RepositoryVM --vmtemplate RepositoryVM_template --vm-size small --use-address-pool Y 8. Verify this was successful and make a note of the IP address of the Repository-VM (in this example it is 10.129.71.30): admin@node003:~$ hspadm vm-instance list You can also check that the repository is running by browsing to its IP address. Deploy the Ambari VMs on HSP 3
9. The Ambari installation process described later in this document utilizes an ssh private key for password-less ssh. You can either: Use the default ssh key which is included in the vm-template. or Go to Deploy the Ambari template. Change the ssh key. Go to Change the ssh key (optional steps) Change the ssh key (optional steps) If you want to change the ssh key, please follow the steps in this section. a. Deploy an Ambari VM instance in sandbox mode. admin@node003:~$ hspadm vm-instance add --name ambari_sandbox -- vm-template ambari_template1 --is-sandbox Y b. Verify this is successful, and make a note of the IP address of the instance (in this example the IP address is 10.129.71.20): admin@node003:~$ hspadm vm-instance list c. Log into the vm sandbox instance. ssh 10.129.71.20 (sandbox instance address) Username: root Password: smrace1 4 Deploy the Ambari VMs on HSP
d. Generate a new ssh key. First, look at the contents of the existing ssh private key. [root@ambari_sandbox ~]# cat.ssh/id_rsa The current ssh key is displayed. Now create the new ssh key. [root@ambari_sandbox ~]# ssh-keygen Output Generating public/private rsa key pair. Enter file in which to save the key (/root/.ssh/id_rsa): /root/.ssh/id_rsa already exists. Overwrite (y/n)? y Enter passphrase (empty for no passphrase): Enter same passphrase again: Your identification has been saved in /root/.ssh/id_rsa. Your public key has been saved in /root/.ssh/id_rsa.pub. The key fingerprint is: The new ssh key is displayed. e. Examine the contents of the new ssh private key to verify that it has changed. [root@ambari_sandbox ~]# cat.ssh/id_rsa You can now use this new ssh key in your Ambari installation. f. Update authorized_keys: [root@ambari_sandbox ~]# cat.ssh/id_rsa.pub >.ssh/authorized_ keys g. From the HSP cluster, power off the VM sandbox instance: admin@node003:~$ hspadm vm-instance shutdown --name ambari_ sandbox h. Go to Deploy the Ambari template below. Deploy the Ambari VMs on HSP 5
Deploy the Ambari template 1. Deploy the vm-template. In this example we are creating a 5 node Hadoop cluster. admin@node003:~$ hspadm vm-instance-group add - name ambaridev -- vm-template ambari_template1 --vm-size medium - num-instances 5 - -use-address-pool Y 2. Verify this was successful: admin@node001:~$ hspadm vm-instance list --template ambari_ template1 Install script package HSP provides a script that automates much of the setup required by Ambari. For example the script sets up the /etc/hosts file so that all of the nodes can find each other. These scripts are packaged as an RPM which can be fetched from the Repository-VM. The node is not configured to use the Repository VM yet so wget is used instead. 1. SSH to any one of the Ambari VM instances, this will become the Ambari server (in this example 10.129.71.21 is used): ssh 10.129.71.21 (VM instance address) Username: root Password: smrace1 2. Fetch the script RPM from the Repository VM (in this example the Repository VM IP address is 10.129.71.30): wget 10.129.71.30/rpm/ambari-2.2.2/AmbariPrep-2.2.2-1.noarch.rpm Verify the RPM was successfully fetched to the current directory. 3. Install the script package using yum: yum -y install AmbariPrep-2.2.2-1.noarch.rpm 6 Deploy the Ambari VMs on HSP
2 Setup Ambari This chapter describes the procedure for setting up Ambari. At the end of this procedure the cluster will be ready to be configured. Setup Ambari 7
1. Log into the first vm-instance. This instance will become the Ambari server. ssh 10.129.71.21 Username: root Password: smrace1 2. Run the install, setup and start of Ambari server. In this example the repository VM IP address is 10.129.71.30, this will enable all of the Ambari VM instances to access the repository: /opt/escale/bin/ambariprep.py --repo-host 10.129.71.30 There are several options that can be given, these can be seen by using the h option. The vm-instances are configured to inherit the domain name that is set on the HSP cluster. Note: the script creates the file system supplied via the -fsname option. The script will fail if the filesystem was created prior. Example: [root@ambaridev-1 ~]# /opt/escale/bin/ambariprep.py -- repo-host <RepositoryVM IP Address> --fsname ambaridevfs --domain gse.hds.com --user admin --pw <admin_password> Upon completion of ambariprep.py, ambari will have been installed, configured and started. Please note that it can take the Ambari server up to 3 minutes to start responding on port 8080 for the GUI. You can follow the progress on ambaridev-1 (i.e. the instance on which you ran ambariprep.py) by: [root@ambaridev-1 ~]# cd /var/log/ambari-server [root@ambaridev-1 ~]# tail f ambari-server.log Wait for text similar to: Example Output 04 May 2016 16:16:16,003 INFO [main] AmbariServer:548 - ********* Started Services ********** 8 Setup Ambari
3 Configure the new cluster This chapter describes the procedure for configuring the new Ambari cluster. At the end of this procedure the Ambari cluster will be configured for use. Configure the new cluster 9
1. Open up a browser on a desktop that has access to the vm-instance running Ambari. This is the Ambari server vm-instance created in the previous section. Connect to: http://10.129.71.21:8080 The login will be: Username = admin Password = admin 10 Configure the new cluster
2. The wizard screen will come up next. Click Launch install Wizard. 3. Assign a name to the new Hadoop cluster. Click Next. Configure the new cluster 11
4. Select the HDP stack version to load on the new cluster. Currently only HDP 2.4 is available with this template. Click Next. 5. Install options. On this page we will need to gather information from the Ambari server. Ssh 10.129.71.21 Username: root Password: smrace1 First we need to gather the FQDNs of each of the vm-instances that will run Ambari agents and be a part of the Hadoop cluster. The ambariprep.py script has already setup the hosts file which is where the information can be pulled from. Example: [root@ambaridev-1 ~]# grep gse /etc/hosts awk '{print $2}' ambaridev-1.gse.hds.com ambaridev-2.gse.hds.com ambaridev-3.gse.hds.com ambaridev-4.gse.hds.com ambaridev-5.gse.hds.com 12 Configure the new cluster
Paste the output into the Target hosts section. Second we will need to get the private ssh key from the Ambari server. This must be pasted into the text box below the radio button indicating Provide your SSH Private Key... Please copy and paste from id_rsa. For example: [root@ambaridev-1 ~]# cat.ssh/id_rsa Output -----BEGIN RSA PRIVATE KEY----- MIIEpQIBAAKCAQEA3qnRvdWW6dlEw3xsj6wtrdtjsAgX3varQaHK/vRWcW84Ijg4 xfqefj7bi19devxt75qrzgq/sofmxexlc4nzlv7z6/xkl920vibhbshfuflx9muy ep7i8ddgtwswoivmr3eeh8un8hxp6btbcuve+cstyjziguw0pgoliepiemdsdz6t poy3zlot5hjydhre2wqezgh7rsyz4/u/hsqivwnkmrv5ad4ike/hrnqxrzslhraa fi2kikarczlpvanqnsygz1/74gefjuhgdzndivsr4ibt9olsmhcgrlyeotifnyny 1FhIU5BOaTCVMwkjxGCfn4nJOES3sRlYMB45gwIDAQABAoIBAQCOhcqgJUBU7OJr 8o+WoxIWD7DOc0+8Qs250/MYoJNYRmV62dvVMW1S1a71BDwwU7WWvYNq/dvZN9yq QW6JxPmEcFJTEAk7Dgkpe72Z2n5PGOew1kUCcAM41UBwP1xZ2ZfqvrdyEoOPEhqo pmlrsonlg/1tvkqqrvoimcq04k7layumgz95ajtjk4s+wkjhykqwolbqmqvbf22w xxllgr1pkwhbgpib346qjshxjkd9urkvy0yas8apjhizerhc+f9bc3ctf/acbzsm kpnm6csx9rio6yed1kfuhwzupueyiu5ezndro9nq5lk4muz8qpwxyf6isnz9dyvs Td6kyKKpAoGBAP7rD2ZNP5gdaIzoLq3bsyaGWBO0g5sJwcwElNyPjN3oR99ZkdXe PQ/2JHPmgXPHksJlhz4KMiqIkZaBirZpt/FyPGHL53Jyo0lC5pITqV682W9nX2G/ xbs0fjiwxfg6pjjzlr5vbrnptp+n6/vtgxesgnshsxl/rcxui1fjgjpfaogban+b t8g17+8qzcnq1om+cmape7m7h1xmnfoo4aydtvttfvwrc+sh9m90oofyn74vc4sf EUcvO1Ndfy6qYjYu3xHeDxOneLIKJyxhLuYW2FX6Kb6HCS8vt2wCBVY/gUhR+igD kiseww2p9j9y4c1gywxns71q8ebdsbb85qp6rs7daogbai9bsfves/1hzjwxls6/ Vyi7NJ0gT0LJYetSSEUhlZ7WGlVJEfKsYLqxOh7T6skp2ybn4f7JzRyOYRG7GD8t hpyludllymtafomebfyigh99joq5v7jbhvoasnfl/gyy8416cgep2zcd24gk9fnq 7GmY4X5Lws5JNo6vlpCwCkERAoGAbB4W1q99u9IcK4jVAvTof2ZdK0jW9d1hGPJm oqzsyipi0tsraxezr1u6y3qkotywblvw7ujhjqxoioaxbjftom5ss8ygb0/xcpvs eurchg/xb1blmi0x2v9/jry/ngbfpgo9or8rpn0w0pvfwlacc9zkqneknczsa0zh GzzUT+ECgYEAyki+BJtUR3sPwBlFgFLBENnvdkcyUFfSHsL77py8tUMBKwxeI+ej 8oF+v/kdPe6+TIneQtzkEBnywqHpSCMyGsPmq2qexICrdaLc/1UxKcRr/1j+i03k ctdbezuxtx5j9rmkjy7546rhwtzsfv8i42aare54roj23xbaxj9immi= -----END RSA PRIVATE KEY----- Important: This is the default ssh private key provided with the VM template. The procedure required to generate a new set of ssh keys is described in Change the ssh key (optional steps). Configure the new cluster 13
6. Click Register and Confirm. 7. All hosts should show success. Click Next. 14 Configure the new cluster
8. Choose Services. Install the following: Yarn + Mapreduce2 Tez Hive HBase Pig Sqoop Zookeeper Ambari Metrics EscaleFS Spark Oozie Falcon Storm Flume Accumulo Atlas Kafka Knox Mahout Slider Currently not in the repo: Configure the new cluster 15
9. Assign Masters. In this step ensure that the Ambari server is only used for nodemanager client. The Ambari server must not run any other services. It is also a good practice to run Hive components (HiveServer, Hive MetaStore) on the same VM. Depending on the load in your environment, you may want to consider running HBase on its own VM. 16 Configure the new cluster
10. Assign Slaves and Clients. Assign Nodemanager for all. RegionServer for all except the ambari server. Phoenix server none. Spark Thrift server none. Client all. Click Next. Configure the new cluster 17
11. Customize Services. There should be no changes required here. Any memory settings can be changed here on a per service basis if needed. A memory warning may come up. This is ok. There is some memory left over and Ambari wants to try and assign 100%. Click Proceed Anyway 18 Configure the new cluster
12. Review. The following information can be saved as a file or printed. This is a high level layout of what is going to be installed. 13. Start Install. Configure the new cluster 19
Everything should complete and present as green. Takes ~10 minutes. Note:The Ambari installation process also runs some smoke tests at this time. You can examine the results by clicking on the Success link. Click Next. 14. Summary of the Install. Click Complete. 20 Configure the new cluster
15. Dashboard view of the up and running Hadoop cluster. Configure the new cluster 21
22 Configure the new cluster
Hitachi Data Systems Corporate Headquarters 2845 Lafayette Street Santa Clara, California 95050-2627 U.S.A. www.hds.com Regional Contact Information Americas +1 408 970 1000 info@hds.com Europe, Middle East, and Africa +44 (0) 1753 618000 info.emea@hds.com Asia Pacific +852 3189 7900 hds.marketing.apac@hds.com MK-95HSP017-03