AutoCollect Process › Setting Up the AutoCollect Environment › Determine the Location of the Source and Repository MUFs

Determine the Location of the Source and Repository MUFs

AutoCollect provides you with the ability to collect performance data from a MUF, known as the source MUF, and store or manipulate it in a different MUF, known as the repository MUF.

• Source MUF communication uses a special navigation process that is only available to Datacom system processes. The snapshot function must be executed on the same LPAR as the source MUF. The snapshot function uses a dynamic navigation process to locate the source MUF and collect the appropriate DST data into memory. The snapshot function only collects one snapshot from one MUF per execution.
• Repository MUF communication uses the standard navigation process. The Snapshot function uses the standard DBSIDPR navigation process to communicate to the repository MUF. This process allows communication to MUFs through SVC (same LPAR), XCF (different LPAR, but same Sysplex), or CCI (different Sysplex). Once communication to the repository MUF is established, the DST information that is collected in memory is written to the snapshot database on the repository MUF as a snapshot rowset.

By using the two forms of communication, AutoCollect can easily access DSTs from one MUF in your environment and store the results as Snapshots on a different MUF in your environment. This is especially helpful when you set up a Complex implementation.

Simple Implementation

You can use the same MUF environment as both the source and the repository for AutoCollect functions. For most cases, this is the typical implementation.

Alternate implementation

At certain sites, such as those with a heavily loaded production system, you may want to limit the overhead of AutoCollect processing on the source environment. In these sites, you can use a different MUF for storing or managing the Snapshot data (repository databases).

Complex implementation

Since AutoCollect attaches the source MUFNAME (required to be unique per installation) to each row that is stored in the repository databases, you can choose to collect data from multiple source MUFs and store it all in one set of repository databases. The AutoCollect programs have been coded to work with this possibility.