State of knowledge

April 2026
ProCall Infinity from version 2603.0 onwards

For this feature to work fully and correctly, both the server and all connected clients must be updated to at least version 2603.0. 
For clients running versions lower than 2603.0, automatic failover is not available or is only available to a limited extent.

Scope of services

The basis and prerequisite is the publicly available documentation for ProCall Infinity (DataCenter) at https://support.estos.de/en/procall-datacenter,
in particular https://support.estos.de/en/procall-datacenter#ProCallDataCenter-Reliabilityandscalability.

Automatic migration of users to available UCServers in the event of a UCServer failure

This document describes the decision-making rules in the login process to enhance the reliabilityand service availability of the UCServer within a ProCall Infinity multi-server environment.

The aim is to intelligently manage user access, minimise the impact of outages and ensure stable telephony operations. 

This feature deals exclusively with UCServer failover and not the reliability of external components (e.g. telephone system, SQL server, network infrastructure).

Measures to maintain stable telephony operations in the event of a UCServer failureWhat impact does a failure have on the overall environment?

If a UCServer fails, this does not mean that the entire multi-server environment is down. The failure affects only the server in question, the clients/users connected to it, and the relevant connections.

Function of automatic user allocation

The following steps ensure that, in the event of a failure, users are automatically and seamlessly redirected to available UCServers without the need for manual intervention by administrators: The remaining UCServer nodes detect the failure via missing health data (Health Monitoring) or connection errors, and the distribution logic is triggered. 

Server logic: Login decision process

The login decision logic evaluates the load (e.g. logged-in users and other metrics) on all nodes and, when a user logs in, determines whether the login should be accepted locally, forwarded to the configured management server, redirected or blocked.

The aim is to ensure an even load distribution, guaranteeing that SIP registrations and TAPI connections are consistently transferred to an active node during relocation; it also regulates the rollback behaviour following the recommissioning of a server.

The rules are designed specifically for scenarios in which a UCServer fails or is unreachable, and specify how logins, SIP registrations and TAPI connections are handled in such cases.

Use lines automatically/dynamically in UCServer

What types of failures and scenarios does this cover?

Specifically, this describes the procedure to follow in the event of a failure of one or more UCServer nodes (e.g. if a UCServer is unreachable) and which  measures to be taken to maintain user logins and telephony services are described.

System requirements 

General system requirements for the ProCall Infinity multi-server architecture

ProCall Infinity multi-server environment in accordance with the ProCall Infinity (DataCenter) system requirements:

ProCall Infinity (DataCenter) version in use

UCServer

ProCall Desktop for Windows 

ProCall Web (Preview)

ECSTA

ECSTA is required when TAPI lines are used.

The UCServer  display all the network available TAPI lines on and not just those on the respective  ECSTA server provided Lines.

For further details on requirements and restrictions, please see: Global line management in ProCall Infinity

Commissioning/Administration

Meeting the system requirements

Before commissioning, ensure that the system requirements are met. 

For ProCall Web (Preview)

Initially assign users to a ProCall Infinity (DataCenter) management server (statically). 
EYou can find instructions on how to do this in the ProCall Infinity (DataCenter) Installation and Configuration Guide

Self-hosting of ProCall Web - Deployment on the customer’s own systems

Configuration settings

Enable/disable the distribution/server logic

The server logic for the automatic distribution of users in the event of a UCServer failure is disabled by default.
To enable the server logic, you must explicitly uncheck the box next to "Disable automatic user distribution" in the UCServer Management under Basic Services > Failover.

Example screenshot: ProCall Infinity UCServer Administration – Basic Services – Restrictions – Fault Tolerance – Disable automatic user distribution


Unforeseen events/Planned measures

Automatic redistribution should generally be enabled in the event of unforeseen events, such as a UCServer failure.

For scheduled tasks, e.g. Wmaintenance work/updates on UCServers, automatic redistribution should be disabled during the maintenance window and re-enabled once the maintenance work has been completed successfully. 

Connections are not automatically migrated when a user’s management server (home server) is changed via UCServer Management.
The only instance where this happens automatically is when a server failure triggers an automatic migration.

Automatic response in the event of a fault: failure of a UCServer

Relocation / User allocation

  • The login logic uses current health scores to select a new management server (home server) for the affected users.
  • Relocation fields (RelocationHomeserverOrigin, RelocationDate) are set in the user profile to indicate the migration.
  • SIP registrations and TAPI/ECSTA connections are transferred to the new node.
  • The client receives a redirect to the backup server.

Bekannte Einschränkung

After a user has been relocated, the client’s login status may be displayed incorrectly in the user list in UCServer Administration. The client is shown as logged in only on the server where the user is currently located, but as logged out on all other servers – even though the client is correctly connected. Reloading ucadmin or restarting the servers does not reliably resolve this display issue.

Return / Recovery

After the system is brought back online, a scheduled recovery job checks the relocation entries and moves users back to their original servers, provided that their original server is healthy and there is no active session.
The relocation back to the original servers is carried out in stages to avoid peak loads.

Recovery generally takes place at 01:00 UTC 

Further information

ProCall Infinity (DataCenter) Release Notes →