Install Hubble Accelerator
This section describes how to prepare your system for Hubble Accelerator and how to install it.
Installation Prerequisites
This section describes how to prepare your system to use Hubble Accelerator. The requirements differ according to the platform on which you want to install Hubble Accelerator.
Software Requirements
Hubble Accelerator is distributed in RPM format and can be installed on the following Linux platforms
- Red Hat Enterprise Linux (64-bit) 7.x
- CentOS 7.x
Dependencies
All dependencies and its dependencies are required. If an offline installation is required, please ensure all dependencies are installed before installing Hubble Accelerator rpm.
- libaio
- perl
- libX11
- libXext
- libXi
- libXrender
- libXtst
- alsa-lib
- perl
- perl-Data-Dumper
- unixODBC-utf16 (For version 4.1.x or higher - shipped with the RPM package)
- initscripts
- atk
- cairo
- gdk-pixbuf2
- gtk2
- pango
- sudo
- which
- mono-complete-6.12.0.107-0.xamarin.9.epel8 (for version 4.2.x or higher)
- bc
- e2fsprogs
- openssl
To install mono-core-6.12.0.107-0.xamarin.9.epel7.x86_64 on CentOS/RHEL 8:
Refer to the detailed instructions for installing Mono on CentOS/RHEL 8 available at https://www.mono-project.com/download/stable/#download-lin-centos
For additional information, you can check the release notes for Mono 6.12.0 at https://www.mono-project.com/docs/about-mono/releases/6.12.0/.
If some dependencies for mono-complete-6.12.0.107-0.xamarin.9.epel7 have been moved from the centos8-stable repository, you can find them at this URL. You can manually download and install them using the following commands:
-
Download the package:
wget <url>
-
Install the package:
yum localinstall <package.rpm>
Uninstallation
To uninstall unixODBC-utf16 for an older version of Accelerator (for example, Version 3.x), you can use the following command:
yum remove -y unixODBC-utf16
To uninstall unixODBC for version 4.2.x or higher, you can use the following command:
yum remove -y unixODBC
To uninstall Mono for version 4.2.x or higher, you can use the following command:
As root user# yum list mono-complete --showduplicateRemove older versions# yum remove -y mono# cd /etc/yum.repos.d/# ls -la centos7-stable.repo# rm centos7-stable.repoDownload centos8-stable repo# wget https://download.mono-project.com/repo/centos8-stable.re# yum install mono-complete
Supported Replication Source Endpoints
Oracle
For supported Oracle DB versions and configurations, refer to Qlik Replicate's documentation.
Oracle Instant Client version 12.1.0.2.0-1.x86_64 will be installed with Hubble Accelerator installation.
Microsoft SQL Server
For supported Microsoft SQL Server DB versions and configurations, refer to Qlik Replicate's documentation.
Microsoft ODBC Driver 18.1 for Windows will be installed on the Qlik Replicate Server machine with Hubble Accelerator installation (for version 4.2.x or higher).
Important: Installing MSODBC 18.1 will uninstall previous versions such as 13.1 and 17.1.
TLS 1.2 prerequisites
If your environment includes:
Microsoft SQL Server 2014
Microsoft ODBC Driver 18 installed on Red Hat 8.1 or later
To enable TLS 1.2, you must install a specific Service Pack on the Windows machine where Microsoft SQL Server is installed. For more details, visit https://support.microsoft.com/en-us/help/3135244/tls-1-2-support-for-microsoftsql-server.
IBM DB2 for z/OS
For supported IBM DB2 for iSeries versions and configurations, refer to Qlik Replicate's documentation.
IBM i Access ODBC 1.1.0.27-1.0 will be installed with Hubble Accelerator installation (for version 4.2.x or higher).
Recommended Host Configuration
Provide a separate disk to hold the Vector database and Qlik replication Data. For example, /etc/fstab:
LABEL=hubble /opt auto defaults 00
Users and Groups
Hubble Accelerator installer will create users and groups in Linux.
Users:
- attunity: Owner of all Attunity related files and directories. This user runs Attunity Replicate on the Accelerator.
- actian: Owner of all Actian related files and directories. This user runs Actian Vector and its related executable on the Accelerator.
- hubble-applianceapi: Owner of Hubble Appliance API related files and directories.
Groups:
-
actian
- members: actian
-
attunity
- members: attunity
-
hubble
- members: hubble-applianceapi
Installation Steps
-
Install the local Hubble Accelerator rpm (replace 4.x.x-x with the actual version of the rpm, example: 4.2.1-29):
yum localinstall hubble-accelerator-4.2.1-29.el8.x86_64.rpm
This step will create the Hubble Accelerator directories and copy the installation packages for all the Accelerator components.
-
Comment
${ACCELERATOR_DIR}/post-install.shscript entry to edit Vector DB parameter settings (Step 4 and 5).cat /opt/insightsoftware/hubble-accelerator/4.2.1/install.sh#!/bin/shset -e# Accelerator installSCRIPT_DIR=$(dirname "$0")ACCELERATOR_PROFILE=/etc/profile.d/accelerator.shACCELERATOR_SOURCE=${SCRIPT_DIR}/.env. ${ACCELERATOR_SOURCE}. ${SCRIPT_DIR}/library.shecho "echo \"Hubble Accelerator: '${ACCELERATOR_VERSION}'\";" > ${ACCELERATOR_PROFILE}# Install Vectorsh ${ACCELERATOR_DIR}/vector/install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# Install Vector patchsh ${ACCELERATOR_DIR}/vector/install-patch.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# Install Vectorsh ${ACCELERATOR_DIR}/vector/install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# Install Attunity Replicationsh ${ACCELERATOR_DIR}/attunity/install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# Install Oracle ODBCsh ${ACCELERATOR_DIR}/oracle/install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# Install MSODBCsh ${ACCELERATOR_DIR}/msodbc/install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# Install IACCESSODBCsh ${ACCELERATOR_DIR}/iaccessodbc/install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# Install Hubble Appliance APIsh ${ACCELERATOR_DIR}/hubble-applianceapi/install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}"# ${ACCELERATOR_DIR}/post-install.sh -p "${ACCELERATOR_PROFILE}" -d "${ACCELERATOR_DIR}" -
Run the Hubble Accelerator installer that will install all the rpms for all the components, and configure them:
Note: Replace 4.x.x with the actual version of the rpm, for example, 4.2.1.
/opt/insightsoftware/hubble-accelerator/3.x.x/install.sh
-
For Accelerator build 4.x.x with Vector version 6.x.x or later, delimited identifiers and case sensitivity database parameter settings need to be catered before creating vector database and users (
Post-install.sh script) during Accelerator installation. Qlik Replicate will create tables and columns with upper or lower cases as per source database metadata and will cause issue with Hubble Application accessing tables/columns with delimited identifiers and upper cases. Make the following changes to update theconfig.datdatabase parameter entry (localhost.createdb.delim_id_case : lower) before creating the default database 'db' and users.Note: Hubble supports table and column names in lowercase.
To set
ii.localhost.createdb.delim_id_caseto lowercase inconfig.dat:-
Go to the directory:
cd $II_SYSTEM/ingres/files/
-
Open
config.dat:vi config.dat
-
Find and change the line to:
ii.localhost.createdb.delim_id_case : lower
-
Save and exit:
:wq!
-
-
For Accelerator build 4.3.x with Vector version 6.3.x or later, Actian has provided the flexibility for switching back to the old behavior of not padding spaces to strings using CHAR / NCHAR functions. When CHAR and NCHAR types are used with x100 tables, the correct behavior is to pad these datatypes with spaces to the explicit and/or implicit length of the datatype. This parameter x100_nopad_char_compat adds a backward compatibility configuration option to allow the user to fall back to the old behavior (Vector version 5.x) where these datatypes were not padded.
To set
ii.localhost.dbms.*.x100_nopad_char_compatto ON inconfig.dat:-
Go to the directory:
cd $II_SYSTEM/ingres/files/
-
Open
config.dat:vi config.dat
-
Find and change the line to:
ii.localhost.dbms.*.x100_nopad_char_compat: ON
-
Save and exit:
:wq!
In Vector version 6.3, the default configuration for string truncation is set to "fail," whereas in Vector version 5.0, it was "ignore."
Update the configuration parameter as follows:
ii.<host>.config.string_truncation: ignore
-
-
After updating the
config.datfile, restart the vector instance and run the post-installation script to implement the non-default parameter. Use the following commands as the actian or root user:Restart the Vector instance (in case of upgrading an Accelerator (for implementing non-default Db parameter) after updating
config.datfile using below commands:systemctl restart actian-vectorVW
or
Restart the vVctor instance and run post installation script (in case of fresh installation of an Accelerator for implementing non-default parameter) after updating config.dat file using below command:
systemctl restart actian-vectorVW Ex: ./opt/insightsoftware/hubble-accelerator/x.x.x/post-install.sh -p /etc/profile.d/accelerator.sh -d /opt/insightsoftware/hubble-accelerator/x.x.x/
Reboot the machine.
-
In case of upgrade, for version 3.3.x or higher, the current vector database 'db' needs to be upgraded:
upgradedb db
-
Test if you can access the Vector database:
sql db
-
Confirm that replication is installed.
repctl version
-
The password for the replication user, and the Hubble user for the database will be in:
cat /opt/insightsoftware/hubble-accelerator/.vector
Feel free to remove the password from that file for security reasons, but leave the file empty.
-
The password for the replication user, in the replication console can be found in:
cat /opt/insightsoftware/hubble-accelerator/.attunity
Feel free to remove the password from that file for security reasons, but leave the file empty.
-
Enable and start the Hubble Appliance API service:
systemctl enable hubble-applianceapi
systemctl start hubble-applianceapi