HCL Workload Automation, Version 9.4

Avoiding full file systems

Perhaps the most important maintenance task to perform is that of regularly controlling the file system or systems where HCL Workload Automation is installed, particularly on the master domain manager.

HCL Workload Automation has a number of files that can grow in size, either with more extensive use, such as the Symphony file, or in the event of network problems, such as the message files. If the Symphony file, in particular, cannot be expanded to contain all the required records, it might become corrupted. If this happens on a fault-tolerant agent or on a domain manager other than the master domain manager, there is a recovery procedure (see the HCL Workload Automation: Troubleshooting Guide). If the Symphony file on the master domain manager is corrupted, you have no alternative but to restart HCL Workload Automation, losing the current plan's workload.

It is thus most important that you monitor the available space on the file system of the master domain manager where the Symphony file is generated, to ensure that there is always sufficient space for it to expand to cover any workload peaks, and also that there is sufficient space for message files to expand in the event of network problems. Your experience with your workload and your network will guide you to determine what are the acceptable limits of available disk space.

The approximate size of the Symphony file can be estimated in advance. It contains items related both to the plan (see Table 1) and to the database (see Table 2). Estimate how many items you have in each category, multiply them by the indicated size in bytes, and sum them to find the approximate Symphony file size:
Table 1. Algorithm for calculating the approximate size of the plan data in the Symphony file
Data in Symphony file from the current plan Bytes per instance
Per Job Scheduler instance: 512
Per job instance: 512
Per job "docommand" string > 40 bytes: The length of the "docommand" string
Per ad hoc prompt: 512
Per file dependency: 512
Per recovery prompt: 512
Per recovery job: 512
Table 2. Algorithm for calculating the approximate size of the database data in the Symphony file
Data in Symphony file from the database (on the master domain manager) Bytes per instance
Per workstation: 512
Per resource: 512
Per user: 256
Per prompt: 512
If the global option ignoreCalendars is set to off, per calendar: 512

If you find that disk space is becoming too limited, and you cannot dynamically extend it, you must create a backup master domain manager with much more space on its file system and then use the switchmgr command so that the backup becomes your new domain manager. Instructions on how to do this for any domain manager are given in Changing a domain manager or dynamic domain manager, and in particular for a master domain manager, in Changing a master domain manager.