|
|
|
|
|
- Required configuration
- Tips
- Creating a cluster
- Configuring the DNS for accessing the cluster
- Configuring the DNS from the HFSQL Control Center
- Installing a new node
- Installing a new HFSQL server on the local computer as cluster node
- Uninstallation
- Uninstalling a node or a cluster
Installing and configuring HFSQL Cluster
Available only with this kind of connection
To install an HFSQL Cluster, you must have: - several computers corresponding to the nodes of the cluster.
- a computer used to host the coordinator (this computer can be one of the cluster nodes).
- a DNS server. Indeed, the access to the cluster is done by sharing the name of a virtual machine associated with the IP addresses of the different nodes.
Reminder A cluster must consist of at least three machines. In order for the coordinator to be installed on the same computer as a node, the cluster must include at least three nodes. If the cluster contains two nodes only, the coordinator must necessarily be found on a third computer. Attention: All servers and workstations must be Windows workstations. The clusters are not available in Linux.
Warning
A specific installation package is required to use HFSQL Cluster. This pack can be downloaded at no extra cost on PC SOFT site, in the download section of modules common to version 2025.
Tips To avoid differences in performance, the configuration of each one of the cluster nodes must have the same order of power (processor and memory). The minimum configuration required to install HFSQL Cluster is as follows (for each node): - Operating system: all Windows server versions supported by the publisher. 32 and 64-bit editions are supported.
- Memory: 512 MB
- Disk: 20 MB for cluster elements + size of databases
- Network: 100 Mbps between nodes. The IP addresses of the nodes must be fixed.
The recommended configuration for installing HFSQL Cluster depends on the size of the databases, their complexity and the number of clients. The following configuration will be robust in most cases (for each node): - Operating system: Windows Server in a version currently supported by the publisher.
It is recommended to use: - Processor: a multi-core processor is recommended
- Memory: 4 GB
- Disk: a RAID disk system
Remarks: - The computer that hosts the coordination service of the cluster requires a less powerful configuration. Typically, the coordinator can be located:
- on one of the nodes,
- on the same server as the DNS of the local network,
- on another computer.
- The port 4997 used by the cluster coordinator must be open in the Windows firewall.
- The ports used by the nodes and the port 4998 must be opened in the Windows firewall on each server.
To create a cluster: - Start the HFSQL Cluster setup wizard.
- Accept the terms in the license agreement to continue the setup.
- Select "Create a new cluster".
- A short text summarizes the principle of a cluster and the steps for its implementation. Go to the next step.
- You can choose to convert an existing HFSQL server into first node of the new cluster or to install a blank cluster. When converting an existing server, specify the name of the server (and the listening port of HFSQL).
- Enter the setup parameters of the coordinator. These parameters are:
- The name of the computer on which the service of the coordinator must be installed.
- The password for managing the cluster. This password will be used in the HFSQL Control Center or in programming for cluster maintenance operations (adding nodes, stopping and restarting, etc.).
- The cluster's DNS name: this is the name from which the cluster will be accessible on the network. For more details, see "Configuring the DNS for the access to the cluster".
- Describe the composition of the cluster. You can add the different nodes of the cluster.
If you have chosen to convert an existing HFSQL server, this one will be automatically proposed. Reminder A cluster must consist of at least three machines. In order for the coordinator to be installed on the same computer as a node, the cluster must include at least three nodes. If the cluster contains two nodes only, the coordinator must necessarily be found on a third computer. - Enter the port number that will be used for the HFSQL connections. This port number will be shared by all the cluster nodes. Therefore, it must be available on each computer that hosts a cluster node. Go to the next step.
New in version 2025Indicate whether you authorize PC SOFT to collect information on the use of the HFSQL Cluster. This data collection is optional, and is used to improve the functionality of the HFSQL Cluster. - The summary of the selected configuration is displayed. You can return to the previous wizard steps to modify the cluster configuration.
- Click "Next". A custom installation package is automatically created. You can:
- Continue setup. The different computers included in the cluster must be accessible.
- Use an installation package. This package must be run on each computer included in the cluster.
Note: Only one coordinator can be installed on a given machine. If two clusters are installed on the same network, their respective coordinators must be found on two different computers. Attention: If you are installing a blank cluster, the following rules must be followed: - never manually copy/paste HFSQL data files into the nodes (even after stopping the nodes).
- always use the cluster to import data:
- via the HFSQL Control Center.
- via an application using the HCopyFile function.
Configuring the DNS for accessing the cluster To give access to the cluster, a record must be added into your DNS server for each node. This operation is automatically performed when installing the cluster and the nodes. When the automatic modification of DNS is not possible (DNS server not supported, rights problem, etc.), the record must be added manually. Remark: To ensure correct operation of the DNS system, it is imperative that the IP addresses of the nodes do not vary.. If the servers get their IP addresses via a DHCP, it must be configured so that a fixed address is reserved for each one of the cluster nodes. Characteristics of the DNS records to create: - Record of (A) type if the nodes are in IPv4 and record of (AAAA) type if they are in IPv6.
- The IP address of the record is the one of the node.
- The name of the record is the DNS name of the cluster. This name can be chosen freely. It must not correspond to the name of a real computer.
When a client application wishes to connect to the cluster, it retrieves the DNS records corresponding to the name chosen for the cluster: this name is the one given as the server address in the Connection variable or in the HOpenConnection or HDescribeConnection functions.. The client computer will select the node used for connection among the different IP addresses. The DNS record can be created once the cluster is installed, by using the HFSQL Control Center. Configuring the DNS from the HFSQL Control Center To configure the DNS from the HFSQL Control Center: - Start the HFSQL Control Center.
- Connect to the cluster by specifying the address of the coordinator.
- If the DNS name of the cluster was not configured yet, a row is displayed in the "Potential problems" table.
- Click the icon for problem resolution (at the end of the row).
- Specify the DNS server to configure, the domain name to use (in most cases, the domain name of the local computer) as well as the authentication parameters if they are required for the connection to the DNS server.
- Validate the window.
- In the dialog box, enter the DNS name that will be used by the client computers to connect to the cluster.
- Validate the dialog box.
All the nodes, except for the first one, must necessarily be new HFSQL servers. You also have the ability to install a new server as first node of the cluster. In this case, the cluster will contain no initial database. Remarks: - Both 32-bit and 64-bit HFSQL servers can be used as nodes of the same cluster.
- All the servers of a cluster must use the same listening port. This port must be chosen during the setup.
Installing a new HFSQL server on the local computer as cluster node To install a new HFSQL server on the local computer as cluster node: - Start the HFSQL Cluster setup wizard.
- Choose "Add a node to an existing cluster".
- Enter the coordinator access parameters: machine name and administration password.
- Modify the composition of the cluster. Several nodes can be added in a single operation.
- A summary allows you to check the cluster configuration before applying the modifications.
- Click "Next". A custom installation package is automatically created. You can:
- Continue setup. The different computers included in the cluster must be accessible.
- Use an installation package. This package must be run on each computer included in the cluster.
Uninstalling a node or a cluster To uninstall a cluster node or the entire cluster: - On the computer of the coordinator, start the HFSQL Cluster setup wizard.
- Select "Uninstall a cluster or nodes".
- Specify the password of the coordinator.
- Check the elements that must be uninstalled.
- Confirm. The uninstall operation is performed.
Remark: This uninstall can also be performed from the HFSQL Control Center.. In this case, you have the ability to temporarily exclude a node from the cluster without having to uninstall it.
This page is also available for…
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|