Prerequisites for HA

To use the High Availability feature, the following items must be provided:

Required ItemPurpose
A Local or Domain User account that has the ability to start and stop locally running services, in particular the PrognosisPostgres service on the High Availability pairs.This account must be used to perform the HA Linking and Switchover plus any other HA related functions.
All the required ports have been opened on the HA Pair, refer to Common Ports in Port Requirements.Ports 1970 and 5432 are needed for Postgres synchronization and communication.
The High Availability product license code (PHA).To run the High Availability software needed for proper operation and for publishing of High Availability records.
Version 11.6 or later deployed on a Management Server and two identical Monitoring Servers (or intermediate Management Servers) to be setup as the High Availability Pair.The latest version of the High Availability feature set is required.

A Security Role for the user of the Administration Tool to perform the setup that has, at a minimum, the following permissions:

  • System (Add and Edit)
  • Configuration
  • Remote Credential
  • Security

To provide the adequate permissions needed to setup, configure and manage the HA Pair.

Network connectivity and remote credentials will need to be set up prior to configuring High Availability.We need administrative access when performing replication changes on the High Availability pair which includes stopping and starting windows services and issuing PostgreSQL database queries and commands.

SECURITY Configuration privileges for the user running the Prognosis service on both the HA Pair servers.

For more information please refer to the SECURITY Configuration Overview.

Allows the Prognosis account to write configuration changes on the High Availability pair when changes occur.

If the Prognosis User is a domain user:

GRANT (*, \MANAGING-NODE, DOMAIN\PROGNOSISUSER)
GRANT (*, \PRIMARY-NODE, DOMAIN\PROGNOSISUSER)
GRANT (*, \SECONDARY-NODE, DOMAIN\PROGNOSISUSER)

If the Prognosis User is NOT a domain user:

On the Management Server
GRANT (*, \MANAGING-NODE, \.PROGNOSISUSER)
GRANT (*, \PRIMARY-NODE, \.PROGNOSISUSER)
GRANT (*, \SECONDARY-NODE, \.PROGNOSISUSER)

On the HA Pair
GRANT (*, \MANAGING-NODE, *)
GRANT (*, \PRIMARY-NODE, \.PROGNOSISUSER)
GRANT (*, \SECONDARY-NODE, \.PROGNOSISUSER)
High Availability Pair servers are reachable/ping-able on the network by referencing their machine name only, without an FQDN.High Availability uses the configured node name to send synchronization status messages. If the machine is not ping-able then the Windows hosts file needs to be modified to resolve the hostname to an IPv4 address from the local machine.
Provide the Prognosis PostgreSQL user with the privileges needed to start and stop the PostgreSQL replication service.

For PostgreSQL to replicate data from the active server to the passive server, the PrognosisPostgres service needs to be stopped on the passive server in order for the PostgreSQL utilities to be executed. Because of this, the pg_replicate script and irpgreplicate executable will need to be run with a user with the privileges to stop and start this service to automatically configure a HA Pair.

The PostgresData folder (i.e. the PostgreSQL database) will also be removed on the newly created passive server the first time that a pair is configured successfully. If this operation fails when the replicate functionality is run we will rollback the original PostgreSQL database folder on the passive server and the pairing will need to be rerun.

For more information about how to setup replication streaming on PostgreSQL, refer to the PostgreSQL documentation on setting up replication: https://wiki.postgresql.org/wiki/Streaming_Replication.

HA has been tested and found to be stable with 10mbps connection speed and 150 ms latency under normal operating conditions.

Ensures that the data transfer rate between the High Availability pair can be consistent and will keep configuration synchronization up-to-date.

If large files and databases with frequently changing contents are configured for synchronization, these requirements are expected to scale accordingly.



Provide feedback on this article