High Availability

This topic is only applicable to on-premises installations.

High Availability (HA) refers to a system or component that is continuously operational in case of hardware failure. Use HA as a preventative measure against down time in the event of failure. To meet HA requirements, Archer can:

  • Deploy on multiple servers
  • Use load balancing technology
  • Utilize shared Unified Naming Convention (UNC)-accessible storage locations
  • Utilize Network Attached Storage (NAS) appliance or clustered file server
  • Create shares on 1 of the participating Services Servers

In addition, you can specify shared, UNC-accessible storage locations for:

  • Archer document repository
  • Keyword search index files
  • Appearance files usually stored in file repositories

These suggestions are recommended to ensure HA performance efficiency:

  • Deploy duplicate web servers running identical services
  • If your environment leverages the Advanced Workflow (AWF) feature, run the AWF service on each web server. AWF uses port 8000 for HTTP and 8443 for HTTPS
  • Configure a load balancer to manage network traffic between web servers

    Note: The Archer Job Engine is self-balancing, and does not rely on a load balancer to distribute workload. Each Job Engine process periodically queries the database and starts the next job in the queue automatically

  • If available, use a preexisting, highly available Network Attached Storage (NAS) appliance or clustered file server. Store the Archer File Repository and Company Files in this shared location
  • Store search indexes locally, as keyword search index performance slows down if search indexes are stored on the NAS. Instead, create shares on 1 of the participating Services Servers

    Note: Most components of Archer failover from 1 server to another automatically when properly load balanced, but flat file search index does not. Manually select a new server to run the search index by its DNS name. This option might not meet the requirements for HA. If you choose this option, scale your server vertically.

  • For the database tier, use an HA-clustered configuration. HA clusters are groups of computers that support server applications, and are reliable with a minimum amount of down-time. It is recommended that you configure HA clusters due to:
    • Industry prevalence
    • Front-end configuration simplicity
    • Fast automatic fail-over

You can also use other enterprise-class HA techniques for database recovery, such as activity groups, log shipping, native mirroring, or regular SQL backup and restore. In most situations, you can independently implement SQL Server HA practices in preexisting environments.

For more information, see "High Availability Multiple Server Environment" in Environment Options.