Why Change to a Database?
When Jobtrac was first designed, BDAM and VSAM were state of the art; the most efficient ways to access data. As your enterprise grew and evolved, you modernized processes and applied state-of-the-art improvements; so did CA. CA Datacom/AD provides the advantages of a relational DBMS, addresses many of your pain points, while actually simplifying the operation of your scheduling solution.
Customer Pain Points
- I need fast and easy access from anywhere - Datacom is much faster and there is full SYSPLEX sharing, as well as a way to share across PLEXes (with some performance penalty.) Clients can use SQL to query the database.
- I need good recoverability and availability - logging and forward recovery is included, as is the ability to perform backup-while-open. A new shadow failover capability is part of R11, significantly enhancing availability and reliability. CAIENF provides recoverable events - Jobtrac will catch up after it comes back up.
- I want scheduling flexibility - R11 includes the ability to have unlimited predecessors and successors; Relational design simplifies access and associations between these, eliminating knitting. You can define multiple Jobtrac tokens in one database.
- I want to keep things simple - using CAIENF common component to track job execution allowed elimination of 3 system exits; this also eliminates the database access required by these exits.
- Please reduce memory demand - the new method of schedule load processing replaces in-core table storage with event staging using database tables. Besides improving memory utilization, this new process serializes adds, eliminating the potential for conflict.
- I need to create customized reports easily - R11 replaces Jreport3 with CA Easytrieve Plus; all standard reports are still provided, but now, these reports are modifiable, or you can create your own from scratch.
Additional Enhancements & Benefits
R11 represents real change and improvement for your Jobtrac solution. In addition to listening to your issues as noted above, the move to CA Datacom/AD provides additional value to your solution.
Checkpoint Read Efficiencies
In the prior release, the master task had to perform sequential reads of the checkpoint file, looking through large volumes of data to find the few records that need processing. This function has been eliminated with R11. Instead, each task reads its own set of records, only focusing on those that require action, such as submission, termination or other functions. The reduction in I/O overhead, in CPU consumption and time is significant. These efficiencies are only possible with the new CA Datacom/AD capabilities.
Speedier Job Submission With R11, submission was completely redesigned, in part to eliminate the use of an IBM macro. This macro tended to caused delays when submitting to an internal reader. Now, job submission is measurably faster.
Simplified Database Management
When you find it necessary to take down the CA Datacom/AD database, the last thing you want to do is stop Jobtrac. Fortunately, this is not necessary. A new and simple suspend/resume process has been added to help simplify the operation.
Automatic SUSPEND and RESUME ensures the high availability you require. This process can be manually invoked through operator commands, or automatically invoked if the MUF (multi-user facility) is shutdown or fails, or if the system fails. In this case, the CA-Jobtrac task will auto-suspend. Immediately, it will attempt reconnect via the auto-resume capability until an instance of CA Datacom/AD is restarted somewhere in the SYSPLEX or a Shadow Datacom/AD assume control. Once the connection is re-established, scheduling will continue just where it left off, with no manual intervention.
The Shadow MUF is a new feature for CA Datacom/AD in R11. It is designed to expedite the recovery of a failed Datacom MUF or system. The Shadow MUF task runs in another LPAR/system of the SYSPLEX, watching the primary MUF through XCF. When it is started, the Shadow MUF does preparatory planning for a failure by opening the same databases and establishing connectivity to the primary. It also creates internal structures needed in the event of an outage. This pre-work is designed to expedite recovery even further.
If the primary fails (system or MUF), CA Jobtrac notes this immediately and auto-suspends and resumes itself. At the same time, the Shadow takes over automatically. CA Jobtrac can then reconnect to the new primary MUF.
- Do I have to be a DBA to deal with this new database?
No, in fact, CA Datacom/AD is delivered as a pre-installed SMP/E environment. CA-Jobtrac's "Planning an Implementation Guide" includes the Datacom install tasks to provide a simpler installation path. CA-Datacom is preconfigured for use in most environments. CAJobtrac provides customized jobs to perform all the needed utilities with the product, and documentation makes it easy to support. Selecting a CA database enabled our developers to make it easier to support and to more closely weave it into Jobtrac than if we had picked another ODBC.
- But I really liked BDAM and VSAM - I don't know anything about Datacom. What now?
The good news is that you don't have to know Datacom to benefit from the new features. All you need to know to continue to operate has been documented for you. The goal is to make life easier, not harder. And the high availability features and usability enhancements should really encourage your quick migration to this new release.
- Does using a database remove some of the Checkpoint files limitations?
Yes. Using a relation database has eliminated many constraints allowing unlimited predecessors and successors, CPM associations, and email template associations. The structure also lays the ground work for eliminating limits on other components in future releases.
- In prior releases of CA-Jobtrac, all users were required to have write access to the CA Jobtrac Checkpoint file. Does this change in r11?
Yes. In r3.5 and before, the checkpoint was accessed by the system exits, and therefore, all users who submitted jobs on the system had to have write access to the Checkpoint. In r11, there is no longer any file or database access from the system exits. This eliminates the security access requirement. Only the CA-Jobtrac started task and ISPF users who use the CA Jobtrac ISPF panels will need to have access to the database.
- I have TRACXnn user exits installed. Will I need to change my exits?
It depends. If you use exits that reference the event record, these will need to reviewed and updated to use the new database record format. Exits (TRACX) 01, 03, 06, 09, 10, 13, 16, 20, 22, 23, and 30 are passed database records. Depending on the functions you are performing in the exit, you may or may not be using the records. Review your exits to be sure that will work with r11. All the remaining exits, including the TRACX26 security exit, will continue to work with out alteration.
- Do my backup and restore utilities change in the new release?
Yes. Back up and restore utility jobs now use the CA Datacom/AD database utility. Sample jobs are provided for the new processes and are discussed in detail in the CA Jobtrac Reports, Maintenance and Recovery Guide. Sysout Capture Utilities have not changed.
- How is Archiving handled in the new release?
The archiving is the same when run under CA-Jobtrac (0ARCHIVE, 0JOBARCH, 0SYSARCH). In batch the module name has been changed to GJTRJARC, though the parameters for the r3.5 module, GJTRMAIV, are still supported. Archived records are now kept in the same database as the active workload to allow reporting of current and archived jobs in one report.
- How does the new database affect CA Jobtrac's Performance?
In benchmarks performed in CA's labs, a test schedule of 1400 jobs completed in ? the time, when compared to CA-Jobtrac r3.5.
- I used to allocate a temporary Checkpoint dataset for forecasting. Does the new release change how forecasting is done?
Yes. During installation, a forecast database is allocated.. In the forecast database, a token is used to separate the various forecasts. A new forecast program, GJTRFCST, is used to load schedules into the forecast database for the desired token. The forecast program greatly improves the forecast process, allowing for selection of multiple SCL members to forecast, forecasting schedules within schedules, and accepting forecast dates in Gregorian and Julian format. The forecasted jobs can be reported using the provided Easytrieve forecast batch report, or viewed through ISPF using the new FSWAP command.
- How is batch reporting handled in the new release? Is JREPORT3 still used?
JREPORT3 is not longer used for batch reporting. In r11, batch reporting is done using the common component version of CA Easytrieve reporting. The provided set of reports has been designed to provide all the reporting capability of JREPORT3, and can be easily modified by providing selection criteria. These reports do not require a knowledge of Easytrieve, but provide simple variables that can be set to select the data for the report. For more advanced users, the provided reports can be used as a basis for creating custom reports that are not available through JREPORT3.
- I had heard that using CA Datacom/AD in a Sysplex, introduced a single point of failure. Is this true?
CA Datacom/AD r11 includes a failover feature, called Shadow MUF, that is designed to immediately take over the databases in the event of a CA Datacom/AD MUF failure or system failure. CA-Jobtrac is designed to take advantage of this feature and will automatically close all connects to the failed MUF and reconnect to the Shadow. In most cases this process occurs without interruption.
- In this release, it was mentioned that some of the system exits have been eliminated. Will this change what we will see in our batch jobs?
Yes, in some cases. Some messages that used to be issued to the jobs log now appear in the CA Jobtrac started task log instead. Also due to the tracking architecture, changes to the Step Footer feature can no longer be provided.
- I have CA-GSS REXX programs that use the JOBTRAC Address environment. Do the database changes affect the data returned by this interface?
No, the @JOBTRAC interface maintains the same format for requesting data and returning data. Rexx programs written to run with previous releases of CA Jobtrac can be used without modification.
- What if I have problems with CA Jobtrac or CA Datacom, who do I open an issue with?
You can always open an issue with CA Jobtrac support. If you are having a problem and you know it is with CA Datacom, opening an issue with them directly may get you a more direct response to your CA Datacom issues. Most commonly, this would concern CA Datacom installation or some of the database utilities. If you are unsure of the problem cause, or if CA Jobtrac is involved, go to CA Jobtrac support for help. You can contact support for either product through SupportConnect.ca.com.