Zone Data Collectors and Election Process
A data collector is an in-memory database that maintains dynamic information about the servers in the zone, such as server loads, session status, published applications, users connected, and license usage. In this article, we’ll learn the process of how zone data collector and election process works.
Data collectors receive incremental data updates and queries from servers within their zone. Data collectors relay information to all other data collectors in the farm. The data collector tracks, for example, which applications are available and how many sessions are running on each server in the zone. The data collector communicates this information to the data store on behalf of the other servers in the farm. By default, in farms that communicate indirectly with the data store, the first server in the farm functions as the data collector.
Farms determine the data collector according to what level the election preference is set for a server. By default, all servers joining the farm are configured as backup data collectors. When the zone’s data collector fails, a data collector election occurs and a backup data collector replaces the failed data collector.
If the data collector fails, existing and incoming sessions connected to other servers in the farm are not affected. The data collector election process begins automatically and a new data collector is elected almost instantaneously. Data collector elections are not dependent on the data store.
New Data Collector Election
When communications fail between data collectors or between a server and its data collector, the process to choose or elect a new data collector begins. For example:
1. The data collector for Zone 1 has an unplanned failure. If the data collector shuts down correctly, it triggers the election process as it shuts down.
2. The servers in Zone 1 detect the data collector failed and start the election process. The server set to the next highest election preference is elected as the new data collector.
3. The member servers start sending their information to the new data collector for Zone 1.
4. The new Zone 1 data collector replicates this information to the other data collectors in the farm.
Sometimes, you might decide to have a dedicated data collector after your farm is in production. In general, if users experience slow connection times due to high CPU utilization on the data collector, consider dedicating a server to act solely as the zone data collector.
Specifying Backup Data Collectors.
When you create a server farm and whenever a new server joins a zone, a server is elected as the data collector for that zone. If the data collector for the zone becomes unavailable, a new data collector is elected for the zone based on a simple ranking of servers in the zone.
Important: A primary domain controller or backup domain controller must not become the data collector for a zone. This situation may arise if XenApp is installed on Windows domain controllers. Citrix does not recommend such installations.