Multiple Coservers on a Single-Node Platform

A configuration of multiple coservers on a single-node platform is best suited to an SMP computer that:

If your configuration consists of a few nodes with lots of CPUs, memory, and disk space, configure multiple coservers per node. Each coserver requires a coserver section in the ONCONFIG file. You can use the onconfig.xps file as a template for a multiple coserver configuration.

Ensure that each coserver can use the disks independently for parallel I/O. Also ensure correct access permissions for dbspaces and files on each coserver. The use of multiple coservers can help to balance the processing load across the available memory and CPU resources. Multiple coservers are also useful for logging-intensive applications or situations in which extremely rapid recovery from any failure is an urgent requirement.

Parallel processing can occur both within coservers and between coservers. To calculate the number of CPUs available to the database server, multiply the NUMCPUVPS parameter by the number of COSERVER parameters listed in the ONCONFIG file.

To run multiple coservers on a single node

Install one copy of the database server on a computer (icecream9 in this example).
To define the new coservers, add a coserver-specific section to your ONCONFIG file, as the following example shows.
```
COSERVER         1,2,3
   NODE      icecream9
   ...
   END
```
Add an entry to your sqlhosts file to enable the new coserver to accept client connections.
For the hostname field of the sqlhosts file, use the same value that you specified in the NODE configuration parameter in your ONCONFIG file.

For the dbservername field, use the coserver name, which is composed of the value of the DBSERVERNAME configuration parameter followed by a period and the number of the new coserver.
```
dbservername.coserver-number 
```
For more information, see The sqlhosts File.
Initialize the database server with the following command:
```
xctl -C oninit -iy
```