About PDS V2 data sets and statuses

Review information about how the PDS V2 data sets function and related terminology.

In persistent data store V2 (PDS V2), the data sets are defined dynamically from within the Tivoli Enterprise Monitoring Server (the monitoring server) or the Tivoli Enterprise Monitoring Agent (the monitoring agent). The number and size of data sets per application is user-configurable, and a wraparound process is used to allocate and delete the data sets when switching from one data set to the next. PDS V2 data sets also support z/OS® data set encryption.

A history data set is active if it is currently being written to by a monitoring server or a monitoring agent. At any given time, update-type operations against tables in a particular application are directed to only one of the data sets available to the data store; this data set is the active data set. All other data sets defined to the data store for the application are referred as being inactive.

Note: In order for any maintenance (such as making a backup copy) to be performed on a data set, the data set must be inactive.

When the active data set runs out of space or at local midnight (depending on your configuration), an automatic data set switch occurs. The data set switch changes the active data set for an application to the reserved data set, which is a data set that has been designated by PDS V2 as the next available data set for the application.

After the switch occurs, in preparation for the next switch, PDS V2 allocates the next reserved data set. If the user-configurable number of data sets for the data store has been allocated and the data sets are full, the oldest data set is removed and replaced with a newly-allocated, empty data set with the same name of the deleted data set. This process of designating a reserved data set in advance allows the opportunity to address problems that might occur when allocating the data set.

Note: If a reserved data set is not available and cannot be allocated, the oldest data set is reused.

PDS V2 data set statuses

A number of terms are used to describe the status of the PDS V2 data sets at any given time. The following list summarizes these terms.

active/inactive: A history data set is active if it is currently being written to by a monitoring server or monitoring agent. All other data sets defined to the data store for the application are referred as being inactive.
empty/partial/full: A data set is empty if no historical data has been written to it. A data set is partial if some historical data has been written to it and there is still space available. A data set is full when there is not enough space for additional historical data to be written to it.
read/write: A data set is in read mode typically when it is allocated to the data store but is inactive. A data set is typically in write mode when it is the active data set.
reserved: A reserved data set is an empty data set that is the target data set when a switch occurs. There is usually one reserved data set for an application, but there could be more when PDS V2 starts for the first time. The purpose of the reserved data set is so that the next switch is successful; if there are issues when allocating the data set, the problem can be addressed before the data set is needed, avoiding possible write failures.