Soft Issues - Managing Oracle Staff
Tasks, Roles and Responsibilities
It is important to define each team's R&R's (Roles and Responsibilities). Setup RACI (Responsible, Accountable, Consulted, Informed) spreadsheets for all multi-team activities like "Commissioning of a new server" or "Create new database".
Also see Roles and Responsibilities.
The change management process should (apart from trying to get everything documented) aim at making sure that only valid and essential changes take place to critical production systems. This usually acts as the first (and at times, only) line of defense against production goof ups.
However, management normally wants all changes to go through a change management approval process. The problem with 'micro' managing changes is that with such a lot of changes going into production, the approval process becomes such a mundane task that changes that actually require a greater degree of scrutiny get passed over as just another change. Other possible outcome is that employees start taking chances and lie about their role when things do go wrong. This defeats the very goal of the change process itself.
- Submit an SOP (Standard Operating Procedure) report for approval to the change control committee. This SOP will cover basic admin tasks like adding datafiles, etc. Write it once, invoke it often.
- Setup different levels of approvals rather than making ALL changes go through ALL approvals. It makes sense to have a single authority for each area - database, storage, backups etc. who has the competency to decide which changes can be passed as regular activities, and which require greater visibility and hence requires further management approval. This will make employees accountable for the 'regular' changes as well.
Root Cause Analysis Reports
When things do go wrong, the employees involved are normally required to complete a RCA (Root Cause Analysis) report to management. This report includes things like:
- Change number
- Downtime resulted (When?)
- Role players (Who took action?)
- Sequence of events that lead to the problem (What went wrong?)
- Tasks to prevent future occurrences
Companies need to have staff on standby to look after critical systems. A standby list needs to be maintained and published so users can easily reference it.
Always have a senior DBA on second call.
What prevent Oracle professionals from performing their tasks optimally:
- Inadequate training
- Old and slow equipment (like old workstations and laptops)
- Meaningless and stretched-out meetings
- Too strict change control regime
Managing Staff Numbers
Some of the difficult issues management needs to address:
- How does one retain critical staff?
- Renumeration/ bonuses
- Team buildings
- How do you know if your staff is over worked or idle most of the time?
- How many databases can one DBA look after?
- How many LOC (Lines of Code) can a developer produce per day?