Home

KB0090

How do I Set Up the Task Server?

Applies to: VC/m 2.x

The Task Server is a background job which runs on a networked M system other than the VC/m system which a user is connected to.

The Task Server is used to perform operations which for any reason cannot be done by the system which the user is connected to. It runs in an M system on the machine where the operations need to be performed. For example, if the user is running VC/m on a UNIX system, the task server can be used to move files on a Windows system. It can also be used in cases where it is necessary to run M routines on a different M implementation.

In order to use this functionality, the task server job (%vczn) must be running in the background, and the physical location addresses must be defined using the remote protocol syntax. The task server uses the global ^%gjtask.

In the areas which are serviced by the task server, all ^%vc* globals must be mapped to the central VC/m system. It is also convenient to map the %vc* routines. The global ^%gjtask should not be mapped to the central VC/m system.

Setting up the Remote Protocol

A physical address which uses the remote protocol is made up of three parts, joined as follows:

protocol://remote_node_specification:local_physical_address

The protocol is 'remote' or 'REMOTE'.

The local_physical_address is the physical address of the location, given in the format which you would use for a physical address on a local machine. This is used to specify the physical address when accessed from the remote machine.

The remote_node_specification is a comma-delimited list which specifies how to access the M system on the remote machine. The format depends on the M implementation.

Remote Node Specification on Caché to Caché Systems

Example: ^YAK^\\yak\d\vcmdata\productsource\vcm\vcmtest,YAK,VCMTEST,\\yak\d\vcmserver\mgr

Piece NumberNameUsage
1directorythe Caché database reference for the database which the job will be run in, made up of ^node_name^directory_path
2systemthe node name of the machine
3namespacethe name of the Caché namespace which the job will be run in, ie one which uses the database specified under 'directory' above
4manager directorythe directory path of the database which the task server will run in

The information in the remote node specification is used in different ways at different times. For example:

  • Remote global access (eg to check whether a routine exists or to manipulate G-type components) uses piece 1.
  • Routine gets and puts use pieces 2 and 4 to communicate with the task server (using global %gjtask on the remote system). The task server then uses piece 3 to switch to the target namespace where it gets the routine source and writes to the ^%vcyt global.

Note: On a Windows system, the directory path much not contain a ':' (colon) character. One way to avoid this is to use a UNC path to the database location, via a share that the Windows system grants to itself.

Note: The database in the 'manager directory' must contain all the routines and globals which are required for the task server process, since it does not have access to any routine or global mappings.

Last reviewed: Nov 22, 2004

Back to VC/m Knowledge Base Articles