Normally, the only way to control the runtime operation of Adabas clients is to modify JCL. This of course assumes that you know where the JCL is for the many hundreds of thousands of JCL pieces that exist in your site; and that you have permissions to access and/or modify them. The disparate nature of JCL often results in the loss of control, as opposed to the gain of it.
Adabas System Coordinator provides advanced control over the runtime processing of Adabas clients; ranging from things like control over activity displays to control of debug processing, for example. These controls are applied independently of JCL at runtime so the administrator can manage the entire Adabas client environment from an online administration tool without needing to access the many JCL objects that exist.
The advanced administrative control of the client runtime environment described above is applied from static configuration that you define online. This is excellent because these controls are applied out of hours as well as when the administrative systems are in use. However, further advanced control is needed for clients that are currently running; they run from the static controls initially but there may be reason to adjust controls dynamically, in an emergency for example. System Coordinator allows you to single out clients in any job (TP, batch, etc) and to alter the controls of individual sessions in your live system.
System Coordinator provides a live on-screen information feed about the current activities of Adabas client jobs and sessions. You are able to see every Adabas client running in your systems through these displays. You can do this from a single display session; your display session does not have to be with the same job as the clients that you are watching! You can see the activities of any Adabas clients you wish from one login.
System Coordinator is a plug-in framework for sibling products such as Fastpath, Vista, SAF Security and Transaction Manager. Part of the framework performs latency management for Adabas client sessions. Latency is where a client session is at rest between terminal interactions, for example. At these times it is important that precious TP system resources reclaimed, such as memory, so they are not tied up while the terminal is temporarily dormant. Terminals are latent during user think-time for example. This latency can last for a few seconds up to many minutes or even hours; this is why careful latency management is so important. System Coordinator has options you can use to allow latency management to be carried out locally within the TP system or out to disk, etc.
As stated previously, latency management can be configured to use disk. This means latent TP sessions become crash-recoverable so that if you use static or dynamic transaction routing to accomplish failover within or across systems the Adabas client sessions managed by System Coordinator become crash-recoverable too!
Part of latency management allows you to set timeout for excessively dormant client sessions. This allows resources to be recovered in a timely fashion.
System Coordinator is aware of the surrounding Adabas network and provides display of the active nodes within it, across the System Coordinator group. In addition, various tasks can be performed on some nodes in the network, such as Adabas databases or Coordinator daemons. Some of these tasks may require the presence of a sibling product but others are provided directly by Coordinator. This allows you to discover the network and it allows you to see the network through the eyes (perspective) of other Coordinators in the network too.
Console messages produced by Coordinator and by sibling products can be shown on the console, as expected. Alternatively, console messages can be redirected to a dataset.
You can set automatic retry with optional pause between retires for various response codes. For example, retry for database down can avoid jobs failing so that they tolerate this and automatically retry repeatedly over aperiod until the database becomes available again. Or, contention for holding records can be retried rather than causing jobs to fail.
You can set automatic diagnostic reporting for various error situations. This is especially useful where a problem arises intermittently or out of hours. The diagnostic reporting happens automatically at the time the error occurs, without human intervention.
Another feature for helping to debug system or application problems allows a unified trace to be recorded for individual sessions. This allows all the sibling products to report their trace information into the same place which give a much more coordinated picture of what is happening at the time the trace is taken.
You can instruct System Coordinator, and siblings, to re-load executable code dynamically. This allows you to introduce fix maintenance without having to stop systems for example.
The Adabas System Coordinator contains three major components client, database and group (represented by a daemon). Depending on the sibling products you use and how your site is configured you may need one, two or all three of these components. There is a fourth online administration tool too.
