Child pages
  • Grouper high availability
Skip to end of metadata
Go to start of metadata

Note, this page describes WS availability and grouper client

  1. If you provision to a highly available LDAP or AD and apps read from that it addresses HA
    1. Apps can use DNSSRV or HAProxy or DNS load balancing or whatever LDAP / AD load balancing
  2. The SQL database that Grouper uses can be clustered somehow
    1. Instituions have used Oracle RAC (active/active), dataguard (readonly), etc
  3. Multiple WS and UI app servers/web servers can be load balanced for HA and high loads
    1. The WS is stateless, any load balancing is fine
    2. The UI is stateful (only for CSRF guard), and needs sticky load balancing.  If a node fails the user will need to start over on another node (unless sessions are clustered across app servers)
  4. The loader process can be run on multiple nodes to improve performance and for fault tolerance
    1. This is Grouper 2.3.0 and above only
  5. You can sync a readonly copy of the database (e.g. offsite) and have readonly WS app servers and web servers handling requests (can use client failover or DNS load balancing or something)
  • No labels