installing platform analytics 9.1.2 for symphony · understanding platform analytics get to know...

Platform AnalyticsVersion 9.1.2for Symphony

Installing

SC14-7569-02

��

NoteBefore using this information and the product it supports, read the information in “Notices” on page 37.

This edition applies to version 9, release 1, modification 2 of Platform Analytics (product number 5725-G84) and toall subsequent releases and modifications until otherwise indicated in new editions.

© Copyright IBM Corporation 2013.US Government Users Restricted Rights – Use, duplication or disclosure restricted by GSA ADP Schedule Contractwith IBM Corp.

Contents

Figures . . . . . . . . . . . . . . . v

Tables . . . . . . . . . . . . . . . vii

Chapter 1. Understanding PlatformAnalytics . . . . . . . . . . . . . . 1Platform Analytics hosts . . . . . . . . . . 1System port usage . . . . . . . . . . . . 1

Analytics system ports . . . . . . . . . . 1System architecture diagram . . . . . . . . . 2

Chapter 2. Preparing to install PlatformAnalytics . . . . . . . . . . . . . . 5Selecting the hosts . . . . . . . . . . . . 5

Selecting the database hosts . . . . . . . . 5Selecting the Analytics node hosts . . . . . . 6Selecting the Analytics server host . . . . . . 7Selecting the Analytics reporting server host . . . 7

Chapter 3. Installation requirements . . 9Obtaining the installation files . . . . . . . . 9Pre-installation checklist . . . . . . . . . . 9

Chapter 4. Installing Platform Analytics 11Installing Database . . . . . . . . . . . . 11

Analytics database checklist . . . . . . . . 11Installing the Analytics database . . . . . . 12

Installing Platform Analytics Server . . . . . . 13Analytics server checklist . . . . . . . . . 13Installing the Analytics server . . . . . . . 13

Installing Platform Analytics Node . . . . . . 14Analytics node checklist . . . . . . . . . 14Installing the Analytics node . . . . . . . 15

Verifying that Platform Analytics is up and running 17

Chapter 5. Configuring and optimizingPlatform Analytics . . . . . . . . . . 21Verifying and configuring the Analytics server. . . 21

Enabling and configuring the remoting server(asynchronous data loading mode only) . . . . 21

Configuring the Analytics node. . . . . . . . 22Stopping redundant PERF services . . . . . 22Enabling and configuring the remoting client(asynchronous data loading mode only) . . . . 23

Optimizing the Analytics server . . . . . . . 23Changing the data retention period . . . . . 23

Optimizing the Analytics node . . . . . . . . 25Increasing JVM memory . . . . . . . . . 25Distributing the Analytics node workload . . . 25

Optimizing the Analytics database. . . . . . . 29

Chapter 6. Uninstalling PlatformAnalytics . . . . . . . . . . . . . . 31Uninstalling the Analytics server . . . . . . . 31Uninstalling the Analytics node. . . . . . . . 31

Chapter 7. Troubleshooting PlatformAnalytics . . . . . . . . . . . . . . 33Some or all Analytics database nodes fail to startdue to a memory error . . . . . . . . . . 33

Starting the Analytics database if all nodes failedto start . . . . . . . . . . . . . . . 33Starting the Analytics database if some nodesfailed to start . . . . . . . . . . . . . 33

Vertica data loader timeout error . . . . . . . 34Third party issues . . . . . . . . . . . . 34Fail to create database due to the status of VerticaNode being DOWN . . . . . . . . . . . 35

Notices . . . . . . . . . . . . . . 37Trademarks . . . . . . . . . . . . . . 39Privacy policy considerations . . . . . . . . 40

© Copyright IBM Corp. 2013 iii

iv Installing Platform Analytics 9.1.2 for Symphony

Figures

1. Platform Analytics Data Collectors forSymphony system architecture with Analyticsnodes in two Symphony clusters . . . . . . 3

© Copyright IBM Corp. 2013 v

vi Installing Platform Analytics 9.1.2 for Symphony

Tables

1. Default Platform Analytics server ports . . . 12. Default Platform Analytics node ports . . . . 23. Default Platform Analytics third-party tool

ports . . . . . . . . . . . . . . . 24. Recommended hardware specifications for the

Analytics database hosts . . . . . . . . 55. Recommended hardware specifications for the

Analytics node hosts . . . . . . . . . . 66. Recommended hardware specifications for the

Analytics server host, not using asynchronousdata loading . . . . . . . . . . . . 7

7. Recommended hardware specifications for theAnalytics server host, using asynchronous dataloading . . . . . . . . . . . . . . 7

8. Hardware configuration requirements forTableau Server . . . . . . . . . . . . 8

9. Recommended hardware specifications for theAnalytics reporting server host . . . . . . 8

10. Data loader categories and correspondingloader controller configuration files . . . . 26

© Copyright IBM Corp. 2013 vii

viii Installing Platform Analytics 9.1.2 for Symphony

Chapter 1. Understanding Platform Analytics

Get to know the components, ports, and architecture of IBM® Platform Analytics.

Platform Analytics hostsThese are the types of Platform Analytics hosts that you can install:

Analytics node

Analytics nodes are hosts on which Platform Analytics Data Collectors forSymphony is installed and which collect operational data from PlatformSymphony clusters. Each node belongs to a cluster from which PlatformAnalytics is to collect data. In each cluster, the Analytics node software isinstalled on a Symphony management host. (For failover support, theAnalytics node must first be installed on the Symphony master host.)

Analytics databaseThe Analytics database runs on one or more hosts and stores the clusteroperational data loaded from the Analytics nodes for later reporting andanalysis.

Analytics server

The Analytics server manages the data that the Analytics nodes collect. Youcan perform all server functions using the Analytics Console in the server.

System port usagePlatform Analytics hosts and components use the following system ports. Youmust ensure that the system ports for Platform Analytics are open for the hostsand components to communicate with one another. You must also ensure that yourweb server port is open.

Analytics system portsPlatform Analytics hosts use the following TCP ports by default. The followingtables describe the configuration files that you must modify to change the defaultsettings and the firewall configurations you must set to allow communicationusing these ports between the various hosts.

Analytics server portsTable 1 lists information about the default Analytics server ports.

Table 1. Default Platform Analytics server ports

Port nameDefaultport number Description

Firewall configuration(all bidirectional) Configuration file

PIAM_PORT 9091 Internal port for thetask scheduler.

not applicable ANALYTICS_TOP\conf\pi.conf

PIEM_PORT 9092 The Analytics eventmanager uses thisport to receive eventsfrom the Analyticsserver and nodes.

Allow connections for thisport from each node hostto the server host.

ANALYTICS_TOP\conf\pi.conf

© Copyright IBM Corp. 2013 1

Table 1. Default Platform Analytics server ports (continued)



Remotingserver port

9093 The Analyticsremoting server usesthis port to receivedata from theremoting node.

Allow connections for thisport from the remotingnode to the server host.

ANALYTICS_TOP\conf\remotingserver.xml

Analytics node portsTable 2 lists information about the default Analytics node ports.

Table 2. Default Platform Analytics node ports



ALC port 4046 The Analytics serveruses this port tomanage the Analyticsloader controller (alc)on the Analytics node.

Allow connections forthis port from the serverhost to each node host.

ANALYTICS_TOP/conf/plc.xml

Third-party tool portsTable 3 lists information about the default Analytics third-party tool ports.

Table 3. Default Platform Analytics third-party tool ports


Firewall configuration(all bidirectional)

Vertica port 5433 This port is required for clientconnectivity, such as JDBC.

Allow connections for this portfrom server and node hosts.

System architecture diagramCluster operations data collected from the Analytics nodes is stored in theAnalytics database for subsequent reporting and analysis by IBM PlatformAnalytics to provide the information you need to make business decisions, such as:v Identify and remove resource bottlenecks that are affecting businessv Plan for capacity and future needs by seeing how capacity is being used and

where capacity is availablev Optimize asset utilization by seeing how your infrastructure and resources are

being usedv Improve productivity and efficiency of your applications, workload costs, and

computing environment

The following diagram shows a sample architecture for a system that monitorsmultiple IBM Platform Symphony clusters.

Figure 1 on page 3 illustrates a system architecture where an Analytics node runson a management host in each of two Symphony clusters.

2 Installing Platform Analytics 9.1.2 for Symphony

Figure 1. Platform Analytics Data Collectors for Symphony system architecture with Analyticsnodes in two Symphony clusters

Chapter 1. Understanding Platform Analytics 3

Chapter 2. Preparing to install Platform Analytics

Before installing Platform Analytics, you will select the hosts that meet the detailedsystem requirements and download the Platform Analytics installation packages.

Selecting the hostsYou must select the appropriate hosts that meet the detailed system requirementsfor the Analytics database hosts, Analytics node hosts, Analytics server host, andAnalytics reporting server host.

Selecting the database hostsFor optimal performance of your production database, the Analytics databasecluster should consist of at least three dedicated multi-core hosts running on ahigh-bandwidth network. Since the Analytics database needs to share a largevolume of data among the database nodes in the database cluster during dataloading or data querying, network bandwidth is an important performancebottleneck for a production database. Therefore, the Analytics database clustershould have a Gigabit Ethernet connection with the Platform Analytics reportingserver and the Analytics node hosts.

Use a 1-10 GB full duplex switch for the private network interface and a VLAN orseparate switch for the public network. The switch used for the private networkshould have sufficient bandwidth to enable 1 GB transfer speeds between any pairof nodes.

The hardware requirements are the same for all the intended database hosts. Referto the Vertica Analytic Database documentation and IBM Platform Analytics ReleaseNotes for the latest list of system requirements and supported operating systemsfor the Analytics database hosts.

Input/output (I/O) performance is important for the operation of a database whilefault tolerance is important to safeguard your data. Using a RAID 01 or 10 systemenables the database host to take advantage of data striping and data mirroring.Data striping allows data to be transferred to multiple hard disks concurrently,which improves I/O performance. Data mirroring means that your database doesnot lose data even if a hard disk fails.

Table 4 describes the optimal configuration of the database server depending onthe size of your cluster. The hardware requirements are the same for each databasehost.

Table 4. Recommended hardware specifications for the Analytics database hosts

Cluster size Number of hosts RAM CPU Local hard disk Network

Medium

(100 - 1000 hosts)

3 16 GB 4 × 2.4GHz 10000 RPMSATA/SCSI/SAS/SSDRAID 01 or 10300 GB

Gigabit Ethernet

Large

(more than 1000hosts)

more than 3 32 GB 8 × 2.4GHz 10000 RPMSATA/SCSI/SAS/SSDRAID 01 or 101 TB

Gigabit Ethernet


Data stripingData striping is the technique of segmenting logically-sequential data, such as asingle file, so that the database can assign segments to multiple physical devices(usually disk drives for RAID storage, or network interfaces for grid-orientedstorage) in a round-robin fashion and thus be read or written concurrently.

Automatic data striping is available in certain RAID devices under software orhardware control, and in file systems of clusters. The following parameters areimportant when improving I/O performance:

Stripe width

The number of parallel stripes that can be written to or read fromsimultaneously. This is the number of disks in the RAID system, and as itincreases, the read/write performance of striped data also increases.

Stripe size

The size of the stripes written to each disk. This may also be referred to asblock size, chunk size, stripe length, or granularity.

You should use a large stripe size of at least 1 MB.

If you are using RAID devices, you should use RAID 10 or RAID 01 because itoffers the best performance of all RAID systems and good fault tolerance.

Selecting the Analytics node hostsWhen selecting a host in the cluster to be a Platform Analytics node, you need toensure that the host is running a supported operating system, and that it meets theminimum hardware requirements. See IBM Platform Analytics Release Notes for thelatest system requirements for the Analytics node host.

The Analytics node must run on a management host in each Platform Symphonycluster from which you want to collect data.

For optimal performance of your Analytics node, the host should be running on ahigh-bandwidth network. Since network bandwidth is an important performancebottleneck for the Analytics nodes, the Analytics node host should have a GigabitEthernet connection with the database host.

If the node is running on a Windows host, you should use the 64-bit version ofWindows because Java™ cannot use more than 1638MB of memory on 32-bitplatforms.

Table 5 describes the optimal hardware configuration of the Analytics node basedon the size of the clusters in which the node resides.

Table 5. Recommended hardware specifications for the Analytics node hosts

Cluster size RAM CPU Local hard disk Network

Medium

(100 - 1000 hosts)

4 GB 2 × 2.4GHz 7200 RPMSATA/SCSI/SAS50 GB

Gigabit Ethernet

Large

(more than 1000 hosts)


Gigabit Ethernet


Selecting the Analytics server hostWhen selecting a host to be the Platform Analytics server, you need to ensure thatthe host is running a supported operating system. See IBM Platform AnalyticsRelease Notes for the latest list of supported operating systems for the Analyticsserver host.

Tip: If you select a host that also meets the Tableau Server system requirements,you can also select the Platform Analytics server host to be the Platform Analyticsreporting server host.

For optimal performance, the Analytics server host should be a dedicatedmulti-core host with sufficient memory and input/output performance. Thenetwork bandwidth between the Analytics server, the database hosts, and theAnalytics nodes is a key performance factor in the Analytics server host.

If you are not using asynchronous data loading mode, the hardware configurationshown in Table 6 should be sufficient.

Table 6. Recommended hardware specifications for the Analytics server host, not using asynchronous data loading

RAM CPU Local hard disk Network

4 GB 4 × 2.4 GHz 7200 RPMSATA/SCSI/SAS50 GB

Gigabit Ethernet

If you are using asynchronous data loading mode, memory is a key performancefactor for the Analytics server host.

You should only use the asynchronous data loading mode for sending data fromthe Analytics node to the database over a slow or unstable network.

Table 7 describes the optimal hardware configuration of the Analytics server if youare using asynchronous data loading, depending on the size of your cluster.

Table 7. Recommended hardware specifications for the Analytics server host, using asynchronous data loading


Medium

(100 - 1000 hosts)


Gigabit Ethernet

Large



Gigabit Ethernet

Note:

Avoid using an NFS disk mount instead of a local hard disk.

Selecting the Analytics reporting server hostWhen selecting a host to be the Analytics reporting server host, you must ensurethat it meets the detailed system requirements for Tableau Server shown in Table 8on page 8.

Chapter 2. Preparing to install Platform Analytics 7

Table 8. Hardware configuration requirements for Tableau Server

Operating system RAM CPU Services User accounts

Windows Server 2003 (SP2 orlater)

Windows Server 2008

Windows Server 2008 R2

4 GB and above Dual-core Do not run InternetInformation Services (IIS) toavoid conflicts with the webserver port 80.

Access to an administratoraccount to install softwareand services.

Access to a user account thatthe service can use(optional).

Tip: If the Analytics server host also meets the Tableau Server systemrequirements, you can select the Analytics server host to also be the PlatformAnalytics reporting server host.

Refer to the Tableau Server documentation or to IBM Platform Analytics ReleaseNotes for the latest list of system requirements and supported operating systemsfor the Tableau Server.

The network bandwidth between the Platform Analytics reporting host and thedatabase cluster may be an important performance bottleneck. Therefore, thePlatform Analytics reporting host should have a Gigabit Ethernet connection withdatabase hosts.

Table 9 describes the optimal hardware configuration of the Platform Analyticsreporting server, depending on the size of your cluster.

Table 9. Recommended hardware specifications for the Analytics reporting server host


Medium

(100 - 1000 hosts)


Gigabit Ethernet

Large



Gigabit Ethernet


Chapter 3. Installation requirements

Before you attempt to install Platform Analytics, you must ensure that you havesatisfied all of the installation requirements.

Obtaining the installation filesObtain the necessary installation files for Platform Analytics, the Analyticsdatabase, and the Tableau Server.v Obtain the following files to install Platform Analytics:

– IBM Platform Analytics server installation package– IBM Platform Analytics node installation package

Choose the package that matches the operating system and systemarchitecture on which you intend to install it.

– IBM Platform Analytics documentation packagev Obtain the following files to install the Analytics database:

– Analytics database installation package– IBM Platform Analytics database schema package

v Obtain the following file to install Tableau Server:– Tableau Server installation package

Note: Obtain Vertica and Tableau keys from [email protected] by providingyour name, your email ID, Platform Analytics version (for example, 9.1.2), PlatformAnalytics edition (For example, Standard Ed), and date of purchase.

Pre-installation checklistUse the following pre-installation checklists to verify that you have satisfied all ofthe installation requirements and are ready to install Platform Analytics.

Important: Do not attempt to install Platform Analytics until you have completedthis checklist. Failure to satisfy all installation requirements may cause problemsduring or after the installation.

Remember that system sizing depends on a number of factors, the most significantinclude:v Number of sessions and tasks per day across all clustersv Number of hosts to be monitoredv Number of running sessions and tasks to be monitoredv Number of concurrent running sessions and tasks to be supportedv Amount of session and task data to be retained

Consider these factors, along with the specifications listed in “Selecting the hosts”on page 5, to help plan your installation.

Check the following items before you install Platform Analytics:v Make sure that the Analytics installation package that you downloaded matches

the operating system and the system architecture on which you intend to installit.


v Verify that you can ping the database hosts from the Analytics node hosts,Analytics server host, and Analytics reporting server host.

v Verify that the Analytics node hosts can ping the Analytics server host, and viceversa.

v Make sure that the Analytics server host and Analytics reporting server hostmeet the hardware configuration requirements. If you plan to install both on thesame system, make sure your hardware supports and withstands bothconfigurations.

v Verify that you have enough memory and disk space for each host.


Chapter 4. Installing Platform Analytics

After preparing your hardware, you can install the Analytics database, theAnalytics server, and the Analytics node. Then, review the items on thepost-installation checklist.

Installing DatabaseThe intended database host (or hosts in the cluster) must meet the PlatformAnalytics and Vertica system requirements.

Analytics database checklist


Manually check the following items for the Analytics database hosts in the cluster:v Verify that the operating system type and version of the hosts meet the Vertica

requirements. All hosts in the database cluster must be running the sameoperating system type and version.

v Verify that the hosts default shell for root and current user is BASH.v Ensure that the hosts have at least 1 GB of RAM per logical processor.v Ensure that the hosts have at least 2 GB of swap space.v Ensure that the disk read-ahead is at least 2048 sectors.v Verify that the configured data and catalog directories are created on all cluster

hosts.v Verify the Python version; only version 2.x, higher than 2.3.4, is supported.v Verify the ssh access configuration.v Verify that NTP is running on startup.v Verify that the SELinux is disabled or set it to permissive mode.v Verify that the Transparent Hugepage is set to madvise or disabled.v Verify that the I/O scheduler is set to deadline or noop.v Verify that the CPU frequency scaling is disabled.v Verify that CPU frequency scaling must be disabled.v Define the loopback address.v Ensure that port 5433 is not in use and that /dev/pts is mounted.

Manually check the following items as described in the Vertica Analytic Databasedocumentation, Installation and Configuration Guide before you install the database:v Make sure that you follow the security setting that is required for the Analytics

database installation, as described in the Vertica documentation.v Make a note of disk storage locations.v Make note of IP addresses of the hosts in the database cluster.

– If you are using a host with multiple network interfaces, use the IP addressthat is assigned to the NIC connected to the other cluster hosts, not the NICthat is used for client connections.


– Dynamically assigned IP addresses are not supported.– Subnet Masks must be the same for all hosts.

v Set the host locale (Language).v Set the default time zone.v Remove nonessential applications.v Disable firewall.v Optionally, run Spread on a separate control network.v Back up any existing database.v Provide root SSH access to all the cluster hosts.

Installing the Analytics databaseThe Analytics database, which consists of one or more database hosts, contains thecluster operations data for reporting and analysis. You must install and configurethe Analytics database before you install Platform Analytics.

You can install the Analytics database on a single host or on a cluster of multiplehosts. Installing the database on multiple hosts improves the performance andreliability of your database.

Before you begin

Important: Make sure you have reviewed and satisfied the installationrequirements in “Pre-installation checklist” on page 9.

For details about installing the database, see the Vertica documentation, Installationand Configuration Guide. Follow the steps that are described in the “Before YouInstall” and “Installing Vertica” chapters.

Procedure1. Log in to the intended database host (or hosts in the cluster) as root.2. Extract the installation package to a shared directory that all the cluster hosts

can access.3. Edit the install.config file and configure all required parameters.

You must configure all the required parameters in the install.config file. Toconfigure any required or optional parameter, uncomment the line and specifythe appropriate value.

4. Run the database installation script:./AnalyticsVerticaInstaller.sh -f install.config

Note: When you configure DB_DATA_PATH, the installer sets owner andgroup of the folder to DBA_USER. So it is not recommended to configureDB_DATA_PATH to folder such as /opt, otherwise, the permissions, owner,group of this folder will be changed.The installer performs the following steps:a. Initializes configurations for the installationb. Installs/repairs SSH keys for the root account on the cluster hostsc. Checks prerequisites on the cluster hosts

Note: If there are any errors in the previous three steps, then aninstall.log file is generated in the installation folder. Fix the errors andrerun the installation script.


d. Installs and configures the Analytics databasee. Creates the database connection for Platform Analytics

Note: As part of the previous step, the database installation script creates anew user for the Vertica database administrator. If the user already exists,the script changes the password of the database administrator to the onespecified by DBA_PASSWORD in the install.config file. This change is done toensure that the password used on a cluster of multiple hosts is uniformacross the hosts.

f. Deploys the Platform Analytics database schema

Note: If there are any errors in the previous three steps, then aninstall.log file is generated in the installation folder. Fix the errors, thencomplete the following steps:1) Delete the /opt/vertica directory, and the files and directories under

your DB_DATA_PATH and DB_CATALOG_PATH paths.2) Rerun the database installation script.

What to do next

If you must update any of the parameters in the database install.config file, seethe description of the dbconfig command in Administering Platform Analytics 9.1.2for Symphony.

Installing Platform Analytics ServerThe intended server host must meet the Platform Analytics system requirements.

Analytics server checklist


Manually check the following items:v The Analytics server meets the detailed system and software requirements.v The Analytics server’s operating system can handle daylight savings time

correctly.v Removed any previous installations of Platform Analytics from your host.v The required ports for Platform Analytics and database hosts are free. If a

firewall exists, all these ports must be open in the firewall. All the required portsare TCP.

v The Analytics database host is properly configured and running.v The Analytics server has access to the Analytics database JDBC driver.

Installing the Analytics serverPerform this task to install the Platform Analytics server on a host.

Before you begin


Chapter 4. Installing Platform Analytics 13

Procedure1. Install and configure Tableau Server. Make sure that the Tableau Server

(tabsvc) service has started.

Note: You will find a separate installer for Tableau Server in the installationpackage. Refer to the Tableau Server documentation if you need moreinstallation details. If you do not have internet and if you want to activateTableau Server offline, follow the Tableau Server steps in this link:http://kb.tableausoftware.com/articles/knowledgebase/offline-activation.

2. Log in to the intended Platform Analytics server as administrator.3. Extract the Platform Analytics server package and edit the install.config file.

Important: Configure all the required parameters in the install.config file. Toconfigure any required or optional parameter, uncomment the line and specifythe appropriate value.

4. Navigate to the top-level installation directory and run the Platform Analyticsserver installation package:

Note: Make sure you have administrator privileges before running this batchfile.AnalyticsServerInstaller.bat -f install.config

Note:

v You can receive notifications of events via email. This means you will bemade aware of potential problems in your cluster without having toconstantly monitor the Platform Analytics Console.

v After enabling email notifications, specify the SMTP mail server and theintended recipient of the email notifications. The intended sender is theaccount that is used to send the email notifications.

5. Verify that the pats service has started:v Windows: ANALYTICS_TOP\bin\perfadmin list

Installing Platform Analytics NodeThe intended node host (or hosts in the cluster) must meet the Platform Analyticssystem requirements.

Analytics node checklistManually check the following items:v The Analytics node host meets the detailed system and software requirements.v The Analytics node host operating system can handle daylight savings time

correctly.v The Analytics node must run on a management host in each Platform Symphony

cluster from which you want to collect data.v The required port for the Platform Analytics hosts is free. If a firewall exists, this

TCP port must be open in the firewall.v The Analytics database host is properly configured and running.v The Analytics node host has access to the Analytics database JDBC driver.


http://kb.tableausoftware.com/articles/knowledgebase/offline-activation

Installing the Analytics nodeInstall the Platform Analytics node package on a management host. You must haveat least one Analytics node host in each cluster that Platform Analytics will handle.

Before you begin


You can install the Analytics node in a Symphony cluster, either with or withoutfailover support.v For failover support, you must install the Analytics node on the Platform

Symphony master host and on one or more management hosts. You can chooseto run the Analytics node on a non-master management host to avoid increasingthe load on the master host.

v Without failover support, you can install the Analytics node on the PlatformSymphony master host or a non-master management host. To avoid increasingthe load on the master host, install the Analytics node on a non-mastermanagement host.

The Symphony shared directory must already exist and be set up according to theinstructions in the Platform Symphony installation documentation. You can use thiscommand to configure the shared directory:egoconfig mghost SHARED_TOP

Installing the Analytics node with failover supportFollow this procedure to install the Analytics node in a Platform Symphony clusterwith failover support. This requires you to install the Analytics node on theSymphony master host and on one or more management hosts in the cluster.

Procedure1. Log in to the Platform Symphony master host.2. Extract the Analytics node package, edit the install.config file, and

configure all required parameters.You must configure all the required parameters in the install.config file. Toconfigure any required or optional parameter, uncomment the line and specifythe appropriate value.

Note: For failover to work successfully, you must configure and install theAnalytics node in the exact same path location on each management host.

3. Run the Analytics node installation script:v On UNIX nodes:

./AnalyticsNodeInstaller.sh -f install.config

Or, for silent installation:./AnalyticsNodeInstaller.sh -f install.config -silent

v On Windows nodes:AnalyticsNodeInstaller.bat -f install.config

Or, for silent installation:AnalyticsNodeInstaller.bat -f install.config -silent

4. Run paconfig on the master host to move the Analytics configuration andwork directories to the SHARED_TOP Symphony shared directory.v UNIX: paconfig mghost pa -f


v Windows: paconfig mghost pa username password -f

5. Repeat step 3 to install the Analytics node on one or more other managementhosts in the cluster.

Note: If you want to configure the Analytics node (for instance, runningdbconfig) on another management host other than the master, first run thefollowing command on that host:paconfig mghost pa

6. Log in to the Platform Management Console and run the post-installationaction to register Platform Analytics as an EGO service. To run thepost-installation action, after installing Platform Symphony and navigating tothe Symphony GUI, follow the instructions in the Cluster Configurationpop-up window. This window appears when you enter your user name andpassword (for example, Admin/Admin).Note: Ensure that you enable the pop-up window function in your browser.

7. For Analytics nodes on UNIX hosts, source the Platform Symphonyenvironment.v For csh or tcsh:

source SOAM_TOP/cshrc.platform

v For sh, ksh, or bash:. SOAM_TOP/profile.platform

8. For Analytics nodes on Windows hosts, run the Symphony Command Promptfrom the Windows Start menu.

9. Verify that the data loaders are running:egosh service view alc

Analytics loader controller (alc) is the data loader controller.10. Verify that there are no errors in any of the data loader log files.

The data loader log files are located in ANALYTICS_TOP/log/dataloader.

What to do next

During normal operations, consider running the Analytics node on a non-mastermanagement host to avoid increasing the load on the master host.

Installing the Analytics node without failover supportFollow this procedure to install the Analytics node on a Platform Symphonymanagement host without failover support.

Procedure1. Log in to the intended Analytics node host. Normally, this should be a

Symphony management host other than the master host.2. Extract the Analytics node package, edit the install.config file, and

configure all required parameters.You must configure all the required parameters in the install.config file. Toconfigure any required or optional parameter, uncomment the line and specifythe appropriate value.

3. Run the Analytics node installation script:v On UNIX nodes:

./AnalyticsNodeInstaller.sh -f install.config

Or, for silent installation:./AnalyticsNodeInstaller.sh -f install.config -silent


v On Windows nodes:AnalyticsNodeInstaller.bat -f install.config

Or, for silent installation:AnalyticsNodeInstaller.bat -f install.config -silent

4. If you are installing the Analytics node on a non-master management host, onthe local management node, copy the SOAM_TOP/gui/conf/post_install/analytics_service.xml and SOAM_TOP/gui/conf/post_install/Register_services_PA.xml files to the SHARED_TOP/gui/conf/post_installdirectory. (Skip this step if you are installing the Analytics node on the masterhost.)

5. If you are installing the Analytics node on a non-master management host,locate the following line in the analytics_service.xml file and modify it, asshown:from:<ego:ResourceRequirement>select(’X86_64’)</ego:ResourceRequirement>

to:<ego:ResourceRequirement>select(HOST_NAME)</ego:ResourceRequirement>

where HOST_NAME is the machine name on which the Analytics node isinstalled.(Skip this step if you are installing the Analytics node on the master host.)

6. Log in to the Platform Management Console and run the post-installationaction to register Platform Analytics as an EGO service.

7. For Analytics nodes on UNIX hosts, source the Platform Symphonyenvironment.v For csh or tcsh:



8. For Analytics nodes on Windows hosts, run the Symphony Command Promptfrom the Windows Start menu.

9. Verify that the data loaders are running:egosh service view alc

Analytics loader controller (alc) is the data loader controller.10. Verify that there are no errors in any of the data loader log files.

The data loader log files are located in: ANALYTICS_TOP/log/dataloader.v UNIX: ANALYTICS_TOP/log/dataloaderv Windows: ANALYTICS_TOP\log\dataloader

Verifying that Platform Analytics is up and runningAfter installation, verify that Platform Analytics is installed successfully andworking properly.

Use the following checklists to verify each part of your Platform Analyticsinstallation.

Verifying the Analytics databasev Ensure that the nodes are up and running.


v Verify that you can create a database user for the creation of Platform Analyticsschema.

v Make sure that the K_SAFE parameter in the create schema file is set to thecorrect value. Use the following command to check the value of the K_SAFEparameter:SELECT current_fault_tolerance FROM system;

v Ensure that the network bandwidth utilization is set properly.

Verifying the Analytics nodev Start the Analytics node after you start the Platform Analytics server.v Wait for 10 minutes after you install the Analytics node and then check the data

loader log files to make sure data has been written and there are no issues foreach node.

v Verify that the database is receiving data. If you have configured remote dataloading, then check the log file to make sure data are written correctly.

v If you plan to upgrade Platform Symphony after installing Platform Analytics:– You must either change the value of the SOAM_VERSION parameter in the

perf.conf file to match the new Symphony version, or reinstall the Analyticsnode after upgrading Symphony.

– If you installed the Analytics node on a Windows host, you must also updatethe value of the SOAM_VERSION environment variable in thePA_TOP/conf/setenv.bat file. For example, if you upgraded to PlatformSymphony 6.1.1, change the following line in the setenv.bat file:set SOAM_VERSION=6.1.1

– If you upgrade to Platform Symphony 6.1.1 after installing PlatformAnalytics, add the following line to the ANALYTICS_TOP/conf/plc_pa_ego.xmlfile to enable the new egoconsumergrouploader data loader (or, if the lineexists, make sure it is set to Enable="true"):<DataLoader Name="egoconsumergrouploader" Interval="300" Enable="true"LoadXML="dataloader/egoconsumergrouploader.xml" />

Verifying the Analytics serverv Verify that you can connect to the Analytics database.v When the server runs two data transformers, check the log files. Also log on to

the Analytics database to verify that aggregate_info has the proper date andproper tables updated.

Note: In AGGREGATE_INFO the NEXT_BEGIN_AGG_TIME is in GMT.v Verify that the report table contains data:

select count(*) from rpt_<data transformer name>_raw;

where <data transformer name> can be, for instance, res_alloc_workload_stats.v After you install the Analytics server and Analytics node, wait two hours and

then check the Cluster tab to see if all hosts are added.v Make sure you enable event notification during the Analytics server installation

and start the server before you start the Analytics node.Check the event_manager_conf table to make sure that your event manager is setto active:event_manager_conf.EVENT_MGR_ACTIVE =’Y’

v Verify that email notification is set up correctly. To specify multiple emailaddresses, enter multiple email addresses separated by a semi-colon (;).


Verifying Tableau Serverv Verify that you can activate the server. If offline activation is required, then

proceed with the on-screen instructions.v If you are not using port 80 during the install configuration, write down the port

number.v Make sure you remember the Administrator user name and password.v When you log in as Administrator, make sure you select Embedded Credentials

on the Maintenance page.v Verify that the tableau service is up and running on the Maintenance page.v Verify that the window service for the Tableau Server is set to Automatic and has

started.

Verifying the Analytics reporting serverv Verify that you have extracted the reports package on the Tableau Server.v Verify that the deployment is successful and the workbooks are under the

projects folder.v Make sure that the parb service starts from Windows services console.


Chapter 5. Configuring and optimizing Platform Analytics

After you install Platform Analytics, you can verify and configure the Analyticsserver, configure the Analytics node, and optimize the Analytics server, node, anddatabase.

Verifying and configuring the Analytics serverUse the IBM Platform Analytics Console to verify and configure the Analyticsserver.

Procedure1. Log in to the Analytics server host.2. Launch the Platform Analytics Console.

v Windows: Start > All Programs > >IBM Corporation > Analytics Server >Analytics Console

3. Click Data Collection Node in the navigation tree and verify that the node isrunning correctly.To view the data loader properties, right-click each loader controller instanceand select Loader Properties.

4. Click Scheduled Tasks in the navigation tree and modify the times that thescheduled tasks are running, if necessary.

5. Click Events in the navigation tree and verify that there are no ERROR or FATALevents.

6. Verify the email notification settings.While in Events, click Action > Notification to open the Event Notificationdialog.

Enabling and configuring the remoting server (asynchronousdata loading mode only)

If you intend to use asynchronous data loading mode, perform this task to enableand configure the remoting server.

About this task


Normally, data loaders perform synchronous data loading, whereby they load datadirectly into the Analytics database. In rare cases where the network connectionbetween the Analytics node and the database host is poor, the data loaders willperform asynchronous data loading. In such cases, the data loaders send data tothe Analytics server, and the server then loads the data into the Analytics database.

Procedure1. Navigate to the ANALYTICS_TOP/conf/ directory.2. Rename the remotingserver.TMPL file to remotingserver.xml.3. Edit the remotingserver.xml file and change the IP address to the IP address of

the Analytics server host.


4. Edit the log4j.properties file and uncomment the RemotingServer line.5. Rename the wsm/pars.TMPL file to wsm/pars.conf.6. Start the remoting server:

perfadmin start pars

7. If you are on a Windows host and want the remoting server to startautomatically when Windows starts, change the startup type of the parsWindow service from Manual to Automatic.

Configuring the Analytics nodeYou can perform the following tasks to configure the Analytics node to work moreefficiently with your cluster.v “Stopping redundant PERF services”v “Enabling and configuring the remoting client (asynchronous data loading mode

only)” on page 23

Stopping redundant PERF servicesPlatform Analytics runs its own PERF services (known as the ALC service)independent of the cluster. Since Platform Symphony clusters also have PERFservices running, you should stop the Symphony PERF services to avoidunnecessary redundancy in your cluster.

Procedure1. Log in to any host in the Symphony cluster:

egosh user logon

2. If you logged in to a UNIX host, source the Platform Symphony environment.v For csh or tcsh:



3. If you logged in to a Windows host, run the Platform Symphony commandprompt from the Start menu.

4. Stop the loader controller (plc), data purger (purger), and the PERF Derbydatabase (derbydb) services:egosh service stop plcegosh service stop purgeregosh service stop derbydb

5. Rename the plc_service.xml loader controller configuration file to another fileextension to remove it from the list of services.

Important: Be sure not to use a file extension of .xml.For example:v Rename EGO_TOP/eservice/esc/conf/services/plc_service.xml to

EGO_TOP/eservice/esc/conf/services/plc_service.xml.bak.v If EGO uses a shared directory, rename SHARED_TOP/eservice/esc/conf/

services/plc_service.xml to SHARED_TOP/eservice/esc/conf/services/plc_service.xml.bak.

6. Similarly, rename the purger_service.xml data purger configuration file andthe derby_service.xml Derby database configuration file to another fileextension to remove them from the list of services.

7. Restart EGO to activate these changes:


egosh ego restart

Enabling and configuring the remoting client (asynchronousdata loading mode only)

If you intend to use the asynchronous data loading mode, perform this task toenable and configure the remoting client.

About this task


Normally, data loaders perform synchronous data loading, whereby they load datadirectly into the Analytics database. In rare cases where the network connectionbetween the Analytics node and the database host is poor, the data loaders willperform asynchronous data loading. In such cases, the data loaders send data tothe Analytics server, and the server then loads the data into the Analytics database.

Procedure1. Log in to the Analytics node host.2. Navigate to the ANALYTICS_TOP/conf directory.3. Rename the remotingclient.TMP file to remotingclient.xml.4. Edit the remotingclient.xml file and change the serverBindAddress and

serverBindPort parameters according to the remoting server configuration.5. Edit the perf.conf file and add the following parameters:

ASYNC_LOADING_ENABLED=YASYNC_LOADING_MODE=Remoting

6. Restart the loader controller service (alc).egosh service stop alcegosh service start alc

Note: If you have to restart plc, then wait several seconds until the STATUSshow "STOPPED" before you run the command perfadmin start plc.

Optimizing the Analytics serverYou can perform the following tasks to modify the Analytics server to enhanceperformance.

About this task

The following tasks are optional.

Changing the data retention periodYou can perform this task to change the data retention period.

About this task

A long data retention period can have a significant impact on performance and thedata volume. You can tailor the data retention period according to your businessrequirements to maximize the performance of your Analytics server.

Chapter 5. Configuring and optimizing Platform Analytics 23

The data purger consists of multiple scheduled tasks(PartitionMaintenanceGroup*), which are enabled by default.

Procedure1. Launch the vsql command line.

a. Navigate to the bin subdirectory of the database installation directory.By default, this is /opt/vertica/bin.

b. Run vsql to connect to the database:./vsql -d database_name -p port -U username -w password

where:v database_name is the name of the databasev port is the TCP port number or the local socket file extension in which the

server is listening for connections. The default is port number 5433.v username is the name of the user with which to connect to the database,

instead of the default user (the database administrator).v password is the password for the database user.

Note: As part of installation, the database script creates a new user for theVertica database administrator. If the user already exists, the password ofthe database administrator is changed to the one specified by DBA_PASSWORDin the install.config file. This change is done to ensure that the passwordused on a cluster of multiple hosts is uniform across the hosts.Alternately, you can run vsql with no options to accept the defaults andspecify the administrator password at the prompt.

2. Examine the current data retention periods of the database tables for PlatformAnalytics.v To examine the retention periods for all Analytics database tables, run the

following from the vsql command line:SELECT TABLE_NAME, DATA_DAYS_RANGEFROM SYS_TABLES_TO_PARTITION;

v To examine the retention periods for a specific database table, run thefollowing from the vsql command line:SELECT TABLE_NAME, DATA_DAYS_RANGEFROM SYS_TABLES_TO_PARTITIONWHERE TABLE_NAME=’table_name’;

where table_name is the name of the table you want to examine.The output displays the name of the table and the corresponding data retentionperiod in days.For example, to view the data retention period for theRESOURCE_METRICS_BUILTIN table, run the following from the vsql commandline:SELECT TABLE_NAME, DATA_DAYS_RANGEFROM SYS_TABLES_TO_PARTITIONWHERE TABLE_NAME=’RESOURCE_METRICS_BUILTIN’;

3. Change the data retention period for the appropriate database tables.a. For each database table to change, run the following from the vsql

command line:UPDATE SYS_TABLES_TO_PARTITIONSET DATA_DAYS_RANGE=’retention_period’WHERE TABLE_NAME='table_name';

where:


v retention_period is the new retention period, in daysv table_name is the name of the table you are changing

b. Commit the changes to the database.Run the following from the vsql command line:COMMIT;

For example, to change the data retention period of theRESOURCE_METRICS_BUILTIN table to 2192 days, run the following from the vsqlcommand line:UPDATE SYS_TABLES_TO_PARTITIONSET DATA_DAYS_RANGE=’2192’WHERE TABLE_NAME=’RESOURCE_METRICS_BUILTIN’;COMMIT;

Optimizing the Analytics nodeYou can perform the following optional tasks to enhance the performance of theAnalytics node.v “Increasing JVM memory”v “Distributing the Analytics node workload”

Increasing JVM memoryYou can perform the following task to increase the amount of memory for the JavaVirtual Machine (JVM).

About this task

If you experience a problem with the JVM running out of memory, you canincrease the JVM memory to at least 2 GB for the Analytics node, instead of usingthe default value.

Procedure1. Log in to the Analytics node host.2. Edit the ANALYTICS_TOP/1.2/etc/plc.sh file.3. Navigate to the following line (toward the bottom of the file):

$JAVA_HOME/bin/java -Xms64m -Xmx1600m -cp ${LCP} ${solarisFlag} \

and change the -Xmx1600m parameter to -Xmx2048m.For example:$JAVA_HOME/bin/java -Xms64m -Xmx2048m -cp ${LCP} ${solarisFlag} \

Distributing the Analytics node workloadYou can improve performance by distributing the Analytics node workload amongmultiple hosts or multiple loader controllers within a host so that each host orloader controller is responsible for a specific type (or types) of data loading.

About this task

Based on the type of data, you can categorize the data loaders into two differenttypes, and enable each node or loader controller to be responsible for only onetype of data loader. Table 10 on page 26 describes the two broad data loadercategories and their corresponding loader controller configuration files.


Table 10. Data loader categories and corresponding loader controller configuration files

Data loader category Loader controller configuration file

Resource management-related, whichincludes:v Allocationsv Hostsv Events

plc_pa_ego.xml

Workload-related, which includes:v Sessionsv Tasks

plc_soam.xml

You can perform the following tasks to distribute the Analytics node workload tomultiple hosts or to multiple loader controllers in one node.

Procedure

Use one of the following procedures to distribute the Analytics node workload:v “Distributing the Analytics node workload to multiple hosts”

Distribute the workload to multiple hosts if your Analytics node is experiencingperformance issues and cannot handle all the workload by itself.

v “Distributing the Analytics node workload to multiple loader controllers in onehost” on page 28Distribute the workload to multiple loader controllers within a host to reduceperformance bottlenecks if your Analytics node host is powerful enough tohandle the workload (for example, the host has at least four cores and at least 6GB of memory).

Distributing the Analytics node workload to multiple hostsA single node might experience performance issues when handling high workloadlevels. To resolve this issue, you can distribute the node workload to multiple hostsso that each host is responsible for a specific type of data loading.

About this task

The following procedure assumes that the Analytics nodes are running in failovermode, which is required to distribute the Analytics node workload to multiplehosts.

Procedure1. Install and configure the Analytics node on another host in the same cluster.

For example, if you plan to have one node for each data loader category, youcould install and configure two Analytics nodes: One to handle EGO-relateddata loaders and one to handle SOAM-related data loaders.

2. For each data loader category, create new loader controllers on each host.Since there are two data loader categories, create two loader controllers on eachhost.In this task, PLC_NAME represents the name of the new loader controller thatyou will create.For example, you can create plc_soam to handle SOAM-related data loaders andplc_pa_ego to handle EGO-related data loaders.a. From the SHARED_TOP/eservice/esc/conf/services directory, copy the

alc.xml file to a new file, one for each new loader controller.


For each new loader controller, name the new file alc_ALC_NAME.xml.For example, copy alc.xml to alc_soam.xml and alc_pa_ego.xml.

b. Edit each new alc_ALC_NAME.xml file and specify the new loader controllername.Navigate to the <ego:Command> and Service_Name parameters and edit thevalues to specify the new loader controller name, as follows:1) In the alc_soam.xml file:

...ServiceName="alcsoam"><ego:Command>ANALYTICS_TOP/1.2/etc/plc.sh -f plc_soam.xml</ego:Command>

2) In the alc_pa_ego.xml file:...ServiceName="alcpaego"><ego:Command>ANALYTICS_TOP/1.2/etc/plc.sh -f plc_pa_ego.xml</ego:Command>

c. From the SHARED_TOP/alc/conf directory, create a new subdirectory for eachnew loader controller.Name each subdirectory PLC_NAME.For example, create new subdirectories named plc_soam and plc_pa_ego.

d. Move the loader controller configuration files corresponding to the dataloader categories from the SHARED_TOP/alc/conf/plc directory to the newsubdirectories corresponding to the new loader controllers that will handleeach category of data loaders, respectively.For example:1) Move plc_soam.xml from SHARED_TOP/alc/conf/plc to

SHARED_TOP/alc/conf/plc_soam.2) Move plc_pa_ego.xml from SHARED_TOP/alc/conf/plc to

SHARED_TOP/alc/conf/plc_pa_ego.e. From the SHARED_TOP/alc/conf directory, copy the plc.xml loader controller

file to a new file, one for each new loader controller.For each new loader controller, name the new file PLC_NAME.xml.For example, copy plc.xml to plc_soam.xml and plc_pa_ego.xml.

f. Edit each new loader controller file and change the Port parameter to a newport number and change the PLCDir parameter to the new loader controllerdirectory.For example:1) In the plc_soam.xml file, change the Port value from 4046 to 4047, and

change the PLCDir value from plc to plc_soam.2) In the plc_pa_ego.xml file, change the Port value from 4046 to 4048, and

change the PLCDir value from plc to plc_pa_ego.g. On each host, from the SHARED_TOP/eservice/esc/conf/services directory,

edit each new alc_ALC_NAME.xml file and change the value of theselect(xxx) parameter to specify the host name on which you want thatloader controller to run.For example, if you want the SOAM-related loader controller to run onhost1_name and the EGO-related loader controller to run on host2_name, thenset the select parameters on each host as:v In the alc_soam.xml file:

select(host1_name)

v In the alc_pa_ego.xml file:select(host2_name)

3. Restart EGO to start the new loader controllers:


egosh ego restart

4. Check the service status:egosh service list

Distributing the Analytics node workload to multiple loadercontrollers in one hostIf your node host is powerful enough to handle the workload (for example, thehost has at least four cores and at least 6 GB of memory), you can perform thistask to reduce performance bottlenecks by creating multiple loader controllers andhaving each loader controller be responsible for a specific type of data loading.This enables each of the multiple cores in your host to control a single loadercontroller and be more efficient in sharing the workload among multiple cores.

Procedure1. For each data loader category, create a new loader controller on the same host.

Since there are two data loader categories, create two loader controllers.In this task, PLC_NAME represents the name of the new loader controller thatyou will create.For example, you can create plc_soam to handle SOAM-related data loaders andplc_pa_ego to handle EGO-related data loaders.a. From the SOAM_TOP/eservice/esc/conf/services directory (or, for failover

mode, the SHARED_TOP/eservice/esc/conf/services directory), copy thealc.xml file to a new file, one for each new loader controller.For each new loader controller, name the new file alc_ALC_NAME.xml.For example, copy alc.xml to alc_soam.xml and alc_pa_ego.xml.

b. Edit each new alc_ALC_NAME.xml file and specify the new loader controllername.Navigate to the <ego:Command> and Service_Name parameters and edit thevalues to specify the new loader controller name, as follows:1) In the alc_soam.xml file:

...ServiceName="alcsoam"><ego:Command>ANALYTICS_TOP/1.2/etc/plc.sh -f plc_soam.xml</ego:Command>

2) In the alc_pa_ego.xml file:...ServiceName="alcpaego"><ego:Command>ANALYTICS_TOP/1.2/etc/plc.sh -f plc_pa_ego.xml</ego:Command>

c. From the ANALYTICS_TOP/conf directory (or, for failover mode, theSHARED_TOP/alc/conf directory), create a new subdirectory for each newloader controller.Name each subdirectory PLC_NAME.For example, create new subdirectories named plc_soam and plc_pa_ego.

d. Move the loader controller configuration files corresponding to the dataloader categories from the ANALYTICS_TOP/conf/plc directory (or, for failovermode, the SHARED_TOP/alc/conf/plc directory) to the new subdirectoriescorresponding to the new loader controllers that will handle each categoryof data loaders, respectively.For example:1) Move plc_soam.xml from ANALYTICS_TOP/conf/plc (or

SHARED_TOP/alc/conf/plc) to ANALYTICS_TOP/conf/plc_soam (orSHARED_TOP/alc/conf/plc_soam).


2) Move plc_pa_ego.xml from ANALYTICS_TOP/conf/plc (orSHARED_TOP/alc/conf/plc) to ANALYTICS_TOP/conf/plc_pa_ego (orSHARED_TOP/alc/conf/plc_pa_ego).

e. From the ANALYTICS_TOP/conf directory (or, for failover mode, theSHARED_TOP/alc/conf directory), copy the plc.xml loader controller file to anew file, one for each new loader controller.For each new loader controller, name the new file PLC_NAME.xml.For example, copy plc.xml to plc_soam.xml and plc_pa_ego.xml.

f. Edit each new loader controller file and change the Port parameter to a newport number and change the PLCDir parameter to the new loader controllerdirectory.For example:1) In the plc_soam.xml file, change the Port value from 4046 to 4047, and

change the PLCDir value from plc to plc_soam.2) In the plc_pa_ego.xml file, change the Port value from 4046 to 4048, and

change the PLCDir value from plc to plc_pa_ego.2. Restart EGO to start the new loader controllers:

egosh ego restart

3. Check the service status:egosh service list

Optimizing the Analytics databaseYou can perform this task to optimize the operation of the Analytics database.

Procedure1. Increase the maximum number of client sessions.

You might need to increase the maximum number of client sessions for thedatabase to communicate with the Platform Analytics nodes. In Vertica, this isthe SESSIONS parameter, which you can change from the vsql command line byusing the SET_CONFIG_PARAMETER function:SELECT SET_CONFIG_PARAMETER('MaxClientSessions', SESSIONS_value);

Each Analytics node needs approximately 30 sessions and the Analytics serverneeds approximately 10 sessions. In addition, each intended user that willconcurrently access the live Platform Analytics reports will also need a clientsession.Therefore, for optimal performance of the database, calculate the optimal valueof the SESSIONS parameter in Analytics as:SESSIONS_value = (# of Analytics nodes) × 30 + 10 + (# of concurrent users browsing live reports)

For example, if you have three Analytics nodes, and will have an extra tenusers concurrently accessing live Platform Analytics reports, increasing themaximum number of client sessions to 110 should be sufficient. From the vsqlcommand line, run the following command:SELECT SET_CONFIG_PARAMETER('MaxClientSessions', 110);

2. Modify the default queue timeout of the tm resource pool to 90 seconds.From the vsql command line, run the following command:ALTER RESOURCE POOL tm queuetimeout 90;

3. To obtain optimal performance, store TEMP and DATA in different storagelocations.


TEMP data is distributed across available storage locations based on availablestorage space. However, DATA can be stored on different storage locations basedon predicted or measured access patterns.

4. For each database host in the cluster, set the block device (blockdev) size.The block device is the physical storage device on the database. Set theblockdev size to obtain optimal performance by running the followingcommands for the drives on which your data directory is located:sudo blockdev --getss drivesudo blockdev --getra drive

For example:sudo blockdev --getss /dev/md0sudo blockdev --getra /dev/md0

By default, the first command should return 512. The second command shouldreturn 2048, which means the readahead parameter is set to 1 MB.If the readahead parameter is set too high, the database host may experience aslow mergeout.

5. (Optional) If the readahead parameter is set too high, set the blockdev size tocorrect this problem:sudo blockdev --setra 2048

To retain this setting every time the host restarts, copy this line into the/etc/rc.local file.


Chapter 6. Uninstalling Platform Analytics

Perform this task if you want to uninstall Platform Analytics from a host. If youwant to uninstall the Analytics database, refer to the Installation Guide for theVertica Analytic Database.

Uninstalling the Analytics serverPerform this task to uninstall the Analytics server from a host.

About this task

This will remove all previously installed Platform Analytics services, reports, andprograms from the host.

Procedure1. Log in to the Analytics server host with the same account that you used to

install the Analytics server.2. Navigate to the ANALYTICS_TOP\uninst directory.3. Run the uninstall script:

uninstaller.bat

Uninstalling the Analytics nodePerform this task to uninstall the Analytics node from a Windows host.

About this task

This will remove all previously installed Platform Analytics programs from thehost. This uninstallation procedure only supports Analytics nodes on Windowshosts.

Procedure1. Log in to the Analytics node host with the same account that you used to

install the Analytics node.2. Run the following command to stop all services:

perfadmin stop all

3. Manually remove all files related to the Analytics node.


Chapter 7. Troubleshooting Platform Analytics

The troubleshooting procedures help you to resolve a variety of common problemscenarios.

Some or all Analytics database nodes fail to start due to a memoryerror

Some or all Analytics database nodes may fail to start and issue the followingerror:Large:Memory(KB) Exceeded: Requested = number, Free = number

This error occurs because of a problem with the Resource Manager in the Analyticsdatabase. To resolve the problem, disable the Resource Manager before running thedatabase, then enable the Resource Manager after the database has started. Theresolution method depends on whether all database nodes failed to start, or onlysome database nodes failed to start.

Starting the Analytics database if all nodes failed to startUse this procedure to start the database in the case where all of the nodes havefailed to start.

Procedure1. Manually disable the Resource Manager on all Analytics database nodes.

Perform the following steps on each host in the database cluster:a. Log in to a host in the database cluster.b. Navigate to the directory containing the vertica.conf file.

The directory is the directory containing the catalogs of the database thatyou want to start. This is the Catalog pathname that you were initiallyprompted to specify when you first created the database.

c. Edit the vertica.conf file and add the following line to the end of the file:EnableResourceManager=0

2. Start the Analytics database on all database nodes.3. Re-enable the Resource Manager:

a. Log in to a host in the database cluster.b. Run the following SQL statement from the vsql command line:

SELECT SET_CONFIG_PARAMETER('EnableResourceManager', '1');

Starting the Analytics database if some nodes failed to startUse this procedure to start the database in the case where only some of the nodeshave failed to start.

Procedure1. Log in to an Analytics database node that is still running.2. Disable the Resource Manager.

Run the following SQL statement from the vsql command line:SELECT SET_CONFIG_PARAMETER('EnableResourceManager', '0');


3. Start the Analytics database on all of the database nodes that failed to start.4. Re-enable the Resource Manager:

a. Log in to a host in the database cluster.b. Run the following SQL statement from the vsql command line:

SELECT SET_CONFIG_PARAMETER('EnableResourceManager', '1');

Vertica data loader timeout errorIf you see this error "Cannot get a connection, pool error Timeout waiting for idleobject" while Vertica database is using virtual IP, use this procedure to fix it:

Procedure1. Shut down the PLC service with the following command:

perfadmin stop plc

2. Change the ${PERF_CONFDIR}/datasource.xml file under the Platform Analyticsnode side. Add the following properties under <ds:Properties /> element:<ds:Property>

<ds:Name>maxIdleds:Name>maxIdle>

<ds:Value>0ds:Value>0</ds:Value>

</ds:Property>

<ds:Property>

<ds:Name>minIdleds</ds:Name>

<ds:Value>0</ds:Value>

</ds:Property>

3. Start the PLC service with the following command:perfadmin start plc

Third party issuesThe following are common issues during Analytics database installation:

Procedurev The readahead size of disk_name is lower than 2048. Solution: As root/sudo, run

the following command:– /sbin/blockdev --setra 2048 disk_name

v NTP is stopped. Solution: As root/sudo, run the following command:– For RedHat and SuSE based systems: /etc/init.d/ntpd start

– For Ubuntu and Debian systems: /etc/init.d/ntp start

v SELinux appears to be enabled and not in permissive mode. Solution: Asroot/sudo, run the following command:– To disable SELinux: setenforce 0

– To change SELinux to permissive mode: setenforce Permissive

v Transparent hugepages is set to always. Solution: As root/sudo, run thefollowing command:– For RedHat systems: echo never > /sys/kernel/mm/

redhat_transparent_hugepage/enabled

– For other systems: echo never > /sys/kernel/mm/redhat_transparent_hugepage/enabled


v Disks do not have deadline or noop I/O scheduling. Solution: As root/sudo, runthe following command– echo deadline>/sys/block/disk_name/queue/scheduler where disk_name is the

disk device name (for example, sda)v CPUs have invalid cpufreq scaling policies: cpus. Solution: As root/sudo, run

the following command– echo performance>/sys/devices/system/cpu/cpu_name/cpufreq/

scaling_governor where cpu_name is the cpu name (for example, cpu0, cpu1).

Fail to create database due to the status of Vertica Node being DOWNUse this procedure create a database in the case where the status of Vertica Node isDOWN. Not deleting the /tmp/4803 directory causes the node to remain in theDOWN state.

Procedure1. Uninstall the Vertica rpm package by running this command:rpm -e

vertica-[vertica version]

2. Delete the /opt/vertica directory, and the files and directories under yourDB_DATA_PATH and DB_CATALOG_PATH paths.

3. Delete the /tmp/4803 directory.4. Rerun the database installation script.

Chapter 7. Troubleshooting Platform Analytics 35

Notices

This information was developed for products and services offered in the U.S.A.

IBM may not offer the products, services, or features discussed in this document inother countries. Consult your local IBM representative for information on theproducts and services currently available in your area. Any reference to an IBMproduct, program, or service is not intended to state or imply that only that IBMproduct, program, or service may be used. Any functionally equivalent product,program, or service that does not infringe any IBM intellectual property right maybe used instead. However, it is the user's responsibility to evaluate and verify theoperation of any non-IBM product, program, or service.

IBM may have patents or pending patent applications covering subject matterdescribed in this document. The furnishing of this document does not grant youany license to these patents. You can send license inquiries, in writing, to:

IBM Director of LicensingIBM CorporationNorth Castle DriveArmonk, NY 10504-1785U.S.A.

For license inquiries regarding double-byte character set (DBCS) information,contact the IBM Intellectual Property Department in your country or sendinquiries, in writing, to:

Intellectual Property LicensingLegal and Intellectual Property LawIBM Japan, Ltd.19-21, Nihonbashi-Hakozakicho, Chuo-kuTokyo 103-8510, Japan

The following paragraph does not apply to the United Kingdom or any othercountry where such provisions are inconsistent with local law:

INTERNATIONAL BUSINESS MACHINES CORPORATION PROVIDES THISPUBLICATION “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHEREXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIEDWARRANTIES OF NON-INFRINGEMENT, MERCHANTABILITY OR FITNESSFOR A PARTICULAR PURPOSE. Some states do not allow disclaimer of express orimplied warranties in certain transactions, therefore, this statement may not applyto you.

This information could include technical inaccuracies or typographical errors.Changes are periodically made to the information herein; these changes will beincorporated in new editions of the publication. IBM may make improvementsand/or changes in the product(s) and/or the program(s) described in thispublication at any time without notice.

Any references in this information to non-IBM Web sites are provided forconvenience only and do not in any manner serve as an endorsement of those Web


sites. The materials at those Web sites are not part of the materials for this IBMproduct and use of those Web sites is at your own risk.

IBM may use or distribute any of the information you supply in any way itbelieves appropriate without incurring any obligation to you.

Licensees of this program who wish to have information about it for the purposeof enabling: (i) the exchange of information between independently createdprograms and other programs (including this one) and (ii) the mutual use of theinformation which has been exchanged, should contact:

IBM CorporationIntellectual Property LawMail Station P3002455 South Road,Poughkeepsie, NY 12601-5400USA

Such information may be available, subject to appropriate terms and conditions,including in some cases, payment of a fee.

The licensed program described in this document and all licensed materialavailable for it are provided by IBM under terms of the IBM Customer Agreement,IBM International Program License Agreement or any equivalent agreementbetween us.

Any performance data contained herein was determined in a controlledenvironment. Therefore, the results obtained in other operating environments mayvary significantly. Some measurements may have been made on development-levelsystems and there is no guarantee that these measurements will be the same ongenerally available systems. Furthermore, some measurement may have beenestimated through extrapolation. Actual results may vary. Users of this documentshould verify the applicable data for their specific environment.

Information concerning non-IBM products was obtained from the suppliers ofthose products, their published announcements or other publicly available sources.IBM has not tested those products and cannot confirm the accuracy ofperformance, compatibility or any other claims related to non-IBM products.Questions on the capabilities of non-IBM products should be addressed to thesuppliers of those products.

All statements regarding IBM's future direction or intent are subject to change orwithdrawal without notice, and represent goals and objectives only.

This information contains examples of data and reports used in daily businessoperations. To illustrate them as completely as possible, the examples include thenames of individuals, companies, brands, and products. All of these names arefictitious and any similarity to the names and addresses used by an actual businessenterprise is entirely coincidental.

COPYRIGHT LICENSE:

This information contains sample application programs in source language, whichillustrates programming techniques on various operating platforms. You may copy,modify, and distribute these sample programs in any form without payment toIBM, for the purposes of developing, using, marketing or distributing application


programs conforming to the application programming interface for the operatingplatform for which the sample programs are written. These examples have notbeen thoroughly tested under all conditions. IBM, therefore, cannot guarantee orimply reliability, serviceability, or function of these programs. The sampleprograms are provided "AS IS", without warranty of any kind. IBM shall not beliable for any damages arising out of your use of the sample programs.

Each copy or any portion of these sample programs or any derivative work, mustinclude a copyright notice as follows:

© (your company name) (year). Portions of this code are derived from IBM Corp.Sample Programs. © Copyright IBM Corp. _enter the year or years_.

If you are viewing this information softcopy, the photographs and colorillustrations may not appear.

TrademarksIBM, the IBM logo, and ibm.com® are trademarks of International BusinessMachines Corp., registered in many jurisdictions worldwide. Other product andservice names might be trademarks of IBM or other companies. A current list ofIBM trademarks is available on the Web at "Copyright and trademark information"at http://www.ibm.com/legal/copytrade.shtml.

LSF®, Platform, and Platform Computing are trademarks or registered trademarksof International Business Machines Corp., registered in many jurisdictionsworldwide.

Intel, Intel logo, Intel Inside, Intel Inside logo, Intel Centrino, Intel Centrino logo,Celeron, Intel Xeon, Intel SpeedStep, Itanium, and Pentium are trademarks orregistered trademarks of Intel Corporation or its subsidiaries in the United Statesand other countries.

Linux is a trademark of Linus Torvalds in the United States, other countries, orboth.

Microsoft, Windows, Windows NT, and the Windows logo are trademarks ofMicrosoft Corporation in the United States, other countries, or both.

UNIX is a registered trademark of The Open Group in the United States and othercountries.

Java and all Java-based trademarks and logos are trademarks orregistered trademarks of Oracle and/or its affiliates.

Other company, product, or service names may be trademarks or service marks ofothers.

Notices 39

http://www.ibm.com/legal/copytrade.shtml

Privacy policy considerationsIBM Software products, including software as a service solutions, (“SoftwareOfferings”) may use cookies or other technologies to collect product usageinformation, to help improve the end user experience, to tailor interactions withthe end user, or for other purposes. In many cases no personally identifiableinformation is collected by the Software Offerings. Some of our Software Offeringscan help enable you to collect personally identifiable information. If this SoftwareOffering uses cookies to collect personally identifiable information, specificinformation about this offering’s use of cookies is set forth below.

This Software Offering does not use cookies or other technologies to collectpersonally identifiable information.

If the configurations deployed for this Software Offering provide you as customerthe ability to collect personally identifiable information from end users via cookiesand other technologies, you should seek your own legal advice about any lawsapplicable to such data collection, including any requirements for notice andconsent.

For more information about the use of various technologies, including cookies, forthese purposes, see IBM’s Privacy Policy at http://www.ibm.com/privacy andIBM’s Online Privacy Statement at http://www.ibm.com/privacy/details thesection entitled “Cookies, Web Beacons and Other Technologies” and the “IBMSoftware Products and Software-as-a-Service Privacy Statement” athttp://www.ibm.com/software/info/product-privacy.


http://www.ibm.com/privacy

http://www.ibm.com/privacy/details

http://www.ibm.com/software/info/product-privacy

��

Printed in USA

SC14-7569-02

installing platform analytics 9.1.2 for symphony · understanding platform analytics get to know...

Documents