Sybase IQ Multiplex Nodes and Server Overview


Looking for something specific?

IQ_SYSTEM_MAIN, IQ_SHARED_TEMP, and IQ user main dbspaces are shared by all multiplex servers, and all servers need access to the same physical file. Data managed by SQL Anywhere is not shared. Each node requires a separate copy of such data.  Each server has its own set of files in IQ_SYSTEM_TEMP and IQ_SYSTEM_MSG.

Note: It is always recommended to create a DEFAULT DB space to store objects and data on and not on the IQ_SYSTEM_MAIN Dbspace.
 

Coordinator Node

Each multiplex configuration requires a coordinator node. When you convert an existing simplex server to multiplex, it becomes the coordinator node.

The coordinator node:

  • Runs read-only and read-write operations against shared IQ objects.
  • Manages IQ main dbspaces.
  • Manipulates local data in SQL Anywhere system and user tables.

Coordinates all read-write operations on shared IQ objects, including:

  • Shared IQ table locking
  • Shared IQ storage management
  • Providing global transaction IDs for read-write transactions involving shared IQ objects
  • Maintaining the global catalog
  • Controls catalog synchronization for secondary servers
  • Performs schema changes on shared IQ store objects
  • Performs schema changes on SQL Anywhere store objects
  • Maintains and cleans up object versions

Secondary NODES.. ie (reader/writer)

One or more secondary nodes may participate in a Sybase IQ multiplex configuration. One secondary node acts as a designated failover node, the first choice node to assume the coordinator role if the current coordinator is unable to continue.

The number of secondary nodes supported depends on the license purchased, as follows:

  • Demo/Trial Edition: Unlimited secondary nodes
  • Small Business Edition: None (multiplex not allowed)
  • Single Application Server Edition: One secondary node
  • Enterprise Edition: Unlimited secondary nodes (license needed for each)

Secondary nodes:

  • Can be either read-only nodes (reader nodes) or read-write nodes (writer nodes).

Writer nodes:

  • Can run read-only and read-write operations against shared IQ objects.
  • Can manipulate local data in temporary and SA base tables.

Reader nodes:

  • Can run read-only operations against shared IQ objects.
  • Can manipulate local data in temporary and SA base tables.