Every system that was ever implemented requires routine chores to be regularly performed to “keep it current and healthy” or to “deal with anticipated events” -and TM1 models are no different.
Some of these chores might include:
- The loading of updated data pools from source systems;
- updating system security;
- closing an accounting period;
- updating conversion (or other) percentages
as well as many others!
Some general guidelines I find helpful:
- Routine tasks or events should never require TM1 administrator access.
- No task or event should result in TM1 logging runtime errors.
- No task or event should result in server lock-outs or outages that last more than an acceptable period of time (typically seconds).
- Routine tasks or regularly required tasks or events should (wherever possible) be scheduled and not require manual initiation or intervention by any individual.
- Routine or events should never require changes to program code.
- All regular or routine tasks and events should be well documented.
In addition, some more systemic advice includes:
- Data should be systematically monitored and key individuals should be automatically alerted of “critical” errors
- The server instance should be regularly recycled -shut down, restarted and backed up.
- Memory to Disk saves should be coordinated and occur on a routine basis that does not impede user access to the application but ensures capture of all data changes.
- Performance bench marks should be taken at key intervals and routinely checked
Of course, these are just some ideas (based upon my experiences), but I think that they make sense – so I hope you find them helpful.
