Default language.

Limited support BMC provides limited support for this version of the product. As a result, BMC no longer accepts comments in this space. If you encounter problems with the product version or the space, contact BMC Support.BMC recommends upgrading to the latest version of the product. To see documentation for that version, see BMC AMI Check for Db2 13.1.

Servers


BMC AMI Utilities servers are jobs automatically initiated by a master job. BMC AMI Utilities invoke servers to expand parallelism and multitasking. When you submit a job, parameters determine the maximum number of servers that the submitted job can initiate. The job creates a task list to identify the work required to complete the job. The servers select work from this list.

Each server can invoke up to 15 subtasks (depending on the job). For some products, you can control the degree of parallelism with respect to these subtasks.

Number of servers

BMC AMI Utilities servers may be automatically initiated when you submit a standard BMC AMI Utilities job. The BMC AMI Utilities job creates the JCL for the servers using skeletons that you set up and declare, specifying the number of each type of server.

Servers are only initiated as needed to reduce processing time and resource usage.

For example, if a job requests three servers, the job initiates with only one server. After initialization, the product re-evaluates the need for more servers and might add one server per minute to a maximum of three servers until processing is complete. If there is not sufficient work to fill all three servers, the job might complete with only one or two servers. 

The number of servers is controlled by the global product parameters SVRDISK and SVRTAPE.

Server job types—batch tasks versus started tasks

BMC AMI Utilities servers can be run in one of the following three modes:

  • Batch jobs 
    Batch jobs are automatically triggered by processing and always have an owner. You can view them easily in SDSF or via other batch management software.

  • Started tasks 
    Started tasks do not have an owner and are run from members of the PROCLIB specified by your data center. Disk and tape jobs can each have their own members for this purpose. Also, started tasks do not need initiators. You determine the PROC name of the started task when you install BMC AMI Utilities.
  • Internal servers
    If +SVRMODE(M) is active, the internal server starts before any external server.

Important

With +SVRMODE(I), all servers will be internal to the Master job.

Servers and data sharing

In some data centers, you might need to run server jobs within specific LPARs. For example, your enterprise might have a three-way parallel sysplex with a data sharing group running within two of them. You can route the server job to either of the two valid LPARs.

If you are running batch jobs, you can use the SVRDISKJC parameter to specify the job class.

For started tasks (STC), servers are started on the system where the master task is running.

If you cannot use the job class, immediately after the JOB statement specify the following line:

/*JOBPARM S=(sys1,sys2 )

The two variables sys1 and sys2 represent the JES member names for the two LPARs in the MAS (JES spool environment).

Important

The IBM SCHENV parameter is supported in server jobs when it is specified in the master job.

For more information about the JOBPARM statement, see the IBM MVS JCL Reference.

Work distribution

BMC AMI Utilities distribute and balance work among servers, whether for one partitioned object or for many table spaces and indexes. The ability to distribute work even at the object level provides for the best possible balance of work, without requiring you to change the job when objects change size.

For efficient use of memory and to reduce the runtime of any one object, the BMC AMI Utilities server locks on to a partitioned table space or a database until that is complete. This enables the server to load a copy of only one database descriptor (DBD) at a time.

Procedurally, when three servers are requested, BMC AMI Utilities start one, wait for a minute, and check to see if there is sufficient work on the queue. If there is sufficient work, BMC AMI Utilities start a second server. This continues until a maximum of three servers are running. If there is not sufficient work, two servers might do all the work. If there is work pending from many databases and two servers are working on one large object, one servers remains for other objects. This distribution of work is dynamic, changing throughout the entire job.

A server might sit idle so long that the master terminates it. If its services are needed later, the master will start up another server to replace it.

Tip

BMC recommends that you put the largest table spaces in a separate job by themselves in the one of the following scenarios:

  • A large partitioned object uses all the servers, preventing smaller objects from starting
  • Many smaller objects are delaying the start of a large table space

This section provides information about the following topics:


 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*