Installation Requirements
Before you install Incorta, ensure your system meets the following requirements.
Operating System
Incorta Analytics 4.0 works on the following operating systems:
- Mac OS
- Centos 7.5, 7.6+
- Redhat 6.9, 7.0, 7.5, 7.6, 8.0
Disk Specifications
Incorta Analytics 4.3 requires the following:
- At least 16 GB of RAM. This number could increase depending on the amount of data.
- Fast storage disk (e.g. SSD).
- Shared storage disk, accessible by all the services within the same cluster (i.e. loader and analytic services) - The shared disk will be used to store data files (if applicable), parquet files, and snapshot files used by the system.
- Java Development Kit (JDK) 8 u144 or 1.8.0.171 and up. Incorta does not support Java SDK 1.8.0.151 through 1.8.0.171.
Database Requirements
Incorta Analytics requires one of the following RDBMS for metadata:
- Derby (packaged for demos and not recommended for production environments)
- MySQL (5.1 or 5.5)
- Oracle 11g and 12g
When you install Incorta Analytics on an Oracle, or MySQL, database, the user needs to be granted certain privileges in order to complete the installation successfully. If these privileges are not granted, users can experience database authorization issues during the installation process. Refer to the following table for the minimum required privileges.
| Oracle | MySQL |
|---|---|
| CONNECT | CONNECT |
| ALTER TABLE | ALTER TABLE |
| CREATE TABLE | CREATE TABLE |
| INSERT | INSERT |
| CREATE INDEX | CREATE INDEX |
| CREATE OR REPLACE SYNONYM | DROP TABLE |
| CREATE OR REPLACE TRIGGER |
Port Requirements
The following table is a list of the ports required by Incorta Analytics at startup. Using these ports will significantly help avoid startup issues. You set up some of these ports during the installation process. Others are fixed values and cannot be changed.
You can find the catalina.properties file and the server.xml file at the path: <INSTALLATION_PATH>/IncortaNode/services/<SERVICE_FOLDER>/conf/.
Other path names in the following table begin with <INSTALLATION_PATH>/IncortaNode/.
| Port Number | Access Type | Purpose | Install Parameter? | Where to change it | Notes | |
|---|---|---|---|---|---|---|
| 4500 | Internal | Node Agent Port | Y | /nodeAgent/nodeAgent.cfg |
Installed on each node to connect that node to the CMC | |
| 8005 | Incoming/ Outgoing | Tomcat shutdown port | Y | /services/ <SERVICE_FOLDER>/conf |
||
| 8080 | Incoming/ Outgoing | Tomcat HTTP port | Y | server.xml |
||
| 8443 | Incoming/ Outgoing | HTTPS traffic | Tomcat HTTPS for SSL | Y | server.xml |
|
| 8009 | Incoming/ Outgoing | Tomcat AJP port | Y | server.xml |
||
| 6060 | Incoming/ Outgoing | HTTP Port | Y | catalina.properties |
||
| 6443 | Incoming/ Outgoing | HTTP/AJP redirect Port | Y | catalina.properties |
||
| 6005 | Incoming/ Outgoing | Server Port | Y | catalina.properties |
||
| 6009 | Incoming/ Outgoing | AJP Port | Y | catalina.properties |
||
| 80 | Incoming/ Outgoing | HTTP/S traffic | Apache Web Server (if installed) | N | Admin Task | If installing Apache, ports 80 and 443 should be open by default. |
| 7777 | Incoming/ Outgoing | Accessed locally by the Incorta Analytics Server | Phantom JS | N | /start-exportserver.sh |
DO NOT CHANGE. |
| 2181 | Internal | Zookeeper | N | /zookeeper/conf/zoo.cfg |
DO NOT CHANGE. | Changing these default ports have a special process. Users need to refer to Zookeeper docs to change them. Or simply avail the ports on the Incorta Analytics Server host. |
| 5442 | Internal | SQL Interface (SQLi) port | N | TMT/Admin UI | If the client (e.g., Tableau) exists on a remote server, this port must be open for remote access. | |
| 5436 | Incoming/ Outgoing | Incorta (memory) | N | TMT/Admin UI | If the client (e.g., Tableau) exists on a remote server, this port must be open for remote access. | |
| 7077 | Incoming/ Outgoing | Spark Master Port | N | Spark configuration | If Spark exists on a remote server, this port must be open for remote access. | |
| 7078 | Incoming/ Outgoing | Spark Worker Port | N | Spark configuration | If Spark exists on a remote server, this port must be open for remote access. | |
| 9091 | Incoming/ Outgoing | Spark Master Web UI Port | N | Spark configuration | For Spark master, this port must be open for remote access to monitor the Spark jobs and logs. | |
| 9092 | Incoming/ Outgoing | Spark Worker Web UI Port | N | Spark configuration | For Spark worker, this port must be open for remote access to monitor the Spark jobs and logs. |
Additional Notes:
- The Shutdown Port 8005 is the shutdown port for the Incorta Analytics Server, and it needs to be accessed by incoming and outgoing data. This port can be configured during the installation process, or changed in the
catalina.propertiesfile. - The HTTP Port 8080 can be configured during the installation process. If it is configured on the Incorta Analytics Server, then it should be made available at startup. You can change the value in the
catalina.propertiesfile. - The HTTPS Port 8443 can be configured during the installation process, or changed in the
server.xmlfile. - The AJP Connector Port 8009 is the Apache Web Server port. It can be configured during the installation process, or changed in the
server.xmlfile.
There are two cases for the Apache Web Server setup:
CASE 1 - If the Apache web server is installed on the Incorta Analytics Server, ensure that the HTTP/S port ”80” (default) has incoming and outgoing data access. Additionally, the AJP connector port “8009” (in this case) must be free internally on the Incorta Analytics Server to communicate between Incorta Analytics and Apache Web Servers. No incoming/outgoing data access is required.
CASE 2 - If the Apache Web Server is NOT installed on the same Incorta Analytics Server, not only must the HTTP/S port ”80” (default) be available and have incoming and outgoing data access, the AJP connector port “8009” must also have an incoming and outgoing data access.
- The Apache Web Server Port 80 is the default port for the Apache Web Server (if installed). It can be changed in the Apache server configuration.
- The PhantomJS 7777 is an internal port that is accessed by the Incorta Analytics Server locally to enable the send-dashboard-by-email feature. No incoming and outgoing data access is required. It CANNOT be set or changed.
- The High Availability (HA) Port 2181 is an internal port that is used as the cluster coordinator for HA environments. It will then require incoming and outgoing data access. It can be set or changed, by changing the property “clientPort = 2181” found in the following path
<INSTALLATION_PATH>/IncortaNode/zookeeper/conf/zoo.cfg
Select a topology
Before you install Incorta Analytics, choose a typical installation for demo or POC setups or a custom installation for HA or cluster setups
Typical Single-Node Topology
Select this installation type for demos and POC setups to install Incorta Analytics locally, with the minimum configuration options. You can only perform one typical installation on the same server. Run this installation to automatically do the following on the server where you are running the installation wizard:
- Install one Cluster Management Console (CMC).
- Install one node.
- Create one cluster called “localCluster”
Post Single-Node Installation
After you finish installing Incorta, you must:
- Log in to the CMC.
- Add at least one loader service and one anaytlics service to the automatically-created cluster “locaCluster”.
Custom Multi-Node Topology
Select this installation type to configure Incorta Analytics for horizontal scalability and to install multiple nodes. A custom installation is highly recommended for deployment in development, UAT, and production environments. For each cluster, run the installation wizard to:
- Install a maximum of one Cluster Management Console (CMC) per cluster (either on a stand-alone server or any of the nodes you plan to use for the cluster).
- Install one node on each server you plan to use for data loading (loader) service or query processing (analytics) service.
Post Multi-Node Installation
After finishing the installation:
- Log in to the CMC.
- Create a cluster.
- Create at least one loader service and one anaytlics service, and attach them to the cluster you created.
- Create a tenant. See Create/Import/Export a tenant using the CMC.