NetVizura needs dedicated server
For security reasons, make sure that your server or VM doesn't have anything installed on it before NetVizura installation. Other software or services running on the same server can impact installation.
NetVizura needs correct time
Before installing NetVizura, make sure to set the time on your server correctly. Time change after the installation will invalidate the license!
NetVizura installation needs internet access
NetVizura requires an active internet connection to install the required dependent software. After the installation is successful, you can disable internet access for the NetVizura server.
Netvizura depends on OpenJDK 8, Tomcat 8.0.14 or higher, and PostgreSQL 12 or higher, and Elasticsearch 8 . NetVizura relies on 3rd-party repositories for installation of these software packages.
The installation process has been tested on Debian 12 and Debian 11. It is important that Debian is a 64-bit OS.
Installation Steps
To install NetVizura, you will need root privileges.
To install NetVizura, follow these steps:
Step 1: Installation of 3rd-party repositories and prerequisite software
Download and execute the Debian prerequisite installation script:
su apt update -y ; apt-get -y install sudo wget wget https://www.netvizura.com/files/products/general/downloads/netvizura-6.1-prerequisites-debian.sh --output-document=/tmp/netvizura-prerequisites-debian.sh bash /tmp/netvizura-prerequisites-debian.sh
Step 2: NetVizura package installation
/tmp directory and execute the following command:sudo dpkg -i /tmp/downloaded_file_name.deb
Now you can go to NetVizura web interface http://<netvizura_server_ip>:8080/netvizura. Default login credentials: For example, if your server IP is 1.1.1.1, then point your browser to http://1.1.1.1:8080/netvizura, like in the screenshot below:
Post Install Steps
After installation, configuration files must be tweaked to utilize the installed RAM to the fullest extent. The main consumers of RAM are the operating system, PostgreSQL/Elasticsearch databases, and Tomcat. The general rule for allocating memory is to split it in a 2:1 ratio between PostgreSQL/Elasticsearch and Tomcat, with 1 GB or more reserved for the operating system. For instance:
| Installed RAM | PostgreSQL | Tomcat | OS | Elasticsearch |
|---|---|---|---|---|
6 GB | 2 GB | 1 GB | 1 GB | 2 GB |
| 16 GB | 2 GB | 5 GB | 1 GB | 5 GB |
Tweaking PostgreSQL
Tweaking PostgreSQL for best performance is a topic on which many books were written, but the following are some common-sense suggestions. For the curious ones, recommended reads (among countless others) are PostgreSQL Optimization Guide and PostgreSQL Tuning Guide.
In order to apply the following tweaks, edit file /etc/postgresql/PG_VERSION_NUMBER/main/postgresql.conf. You will need to restart the PostgreSQL service after editing with the command: systemctl restart postgresql. Almost all of the following parameters are commented with a carron character (#). Be aware that if you comment out the changed parameter, PostgreSQL will revert to the default value.
In the following example, it is assumed that 4 GB of RAM is allocated for PostgreSQL.
Before changing any parameters in PostgreSQL configuration, read the provided comments in the table below for more information regarding a specific parameter.
| parameter | recommended value | comment |
|---|---|---|
max_connections | 30 | NetVizura rarely uses more than 10 connections simultaneously, but it is good to have some reserve. |
shared_buffers | 1024MB | The recommended amount is RAM/4. |
effective_cache_size | 2048MB | The recommended amount is |
checkpoint_completion_target | 0.7 | This parameter can take values between 0 and 1. The default is 0.5, meaning the write phase of the checkpoint process will take half of the checkpoint timeout. Increasing this value will give the checkpoint write phase more time to complete, thereby reducing I/O usage. |
work_mem | 32-64MB | The formula used is |
maintenance_work_mem | 256MB | Speeds up DB self clean process. Usually 4*work_mem or something in that ballpark |
wal_buffers | 16MB | Increasing |
| 1GB | If WAL files are under this size, files will be recycled for future checkpoints. |
| 2GB | The maximum size of WAL files, after which the CHECKPOINT command is issued, and files are written to disk. |
| 2 | The number of simultaneous requests that can be handled efficiently by the disk subsystem. |
full_page_writes | off | Turning this parameter off speeds up normal operation but might lead to unrecoverable or silent data corruption after a power outage, OS failure, or HDD failure. The risks are similar to turning off fsync, though smaller. |
fsync | off | Don't wait for the HDD to finish the previous write operation. This brings the most benefit, but if there is a power outage, an OS or HDD failure at the exact instant when PSQL issues a write command to the HDD, that data will be lost, and the DB itself could be corrupted. On the other hand, DB can issue several times as many write commands in the same time period and consider them all done, thus improving write performance immensely. |
synchronous_commit | off | Similar to "fsync" but with less benefit. |
| Parallel system optimization (PSQL => 9.6) | ||
max_worker_processes | 2 | Number of cores |
max_parallel_workers_per_gather | 1 | Number of cores/2 |
(PSQL > 9.6) max_paralllel_workers | 2 | Number of cores |
Tomcat Memory Allocation
During installation, NetVizura automatically allocates memory for the Tomcat process. The amount allocated to the Tomcat process is calculated according to the formula:
(RAMtotal - 1GB) / 3 but no less than 1GB.
For instance:
| Total RAM | Tomcat |
|---|---|
| 3 GB | 1 GB |
| 4 GB | 1 GB |
| 16 GB | 5 GB |
However, if you need to tweak Tomcat RAM allocation differently (the example for 2048MB):
- Edit file
/etc/default/tomcat8(Debian 9) or/etc/default/tomcat9(Debian 10) - Locate
JAVA_OPTSenvironment variable that defines memory, and uncomment it if it is commented. This line looks something like the following:JAVA_OPTS="${JAVA_OPTS} -Xmx1024m -Xms1024m +UseConcMarkSweepGC" - Modify the
-Xmxparameter to allocate additional memory to Tomcat. Additionally, set parameter-Xmsto the same amount. This should look something like:JAVA_OPTS="-Djava.awt.headless=true -Xmx2048M -Xms2048M -XX:+UseConcMarkSweepGC"
Save the file and restart Tomcat: systemctl restart tomcat8 (Debian 9) systemctl restart tomcat9 (Debian 10)
Elasticsearch Memory Optimization
By default, the memory limit for Elasticsearch should be set to 30% of RAM. If you need it to be set to any other value, edit the file: /etc/elasticsearch/jvm.options, and set values -Xms and Xmx to desired size. Then, restart the Elasticsearch and Tomcat services.