Database Administration and the Goal of Diminishing Downtime

As a DBA, minimizing downtime is one of the most critical parts of day to day life. If your database is down, even for a few minutes, this can translate into lost customers and lost revenue. With the ever increasing shift into online businesses that can be accessed at anytime from anywhere in the world, downtime is something that a DBA can ill afford to let happen.

From a business perspective, it could be argued that you want to sell to your customers 24×7 all year around. This surely would maximize profits and bring on the fastest growth, but at what cost? From a DBA’s perspective, having the highest level of availability may not be within budget. So a compromise may need to be met and it is the responsibility of the DBA to show how much availability can be provided with the budget and tools they have at their disposal. Once this has been laid out for all to see, the DBA must then ensure that the availability that the database has been agreed to is, within reason, and also adhered to.

So with this in mind, what can a DBA do? Downtime and outages are sadly unavoidable for most systems. From a DBA’s perspective, Health Checks and Baselines can help a DBA understand how a system is performing, and pick up on potential resource and storage constraints that may halt a system, as well as pointing to potential tuning opportunities to keep things running at their best, but what about issues outside of a DBA’s control? Having a solid backup and recovery plan that has been tested thoroughly, and can rebuild your system at a moment’s notice, is a must have for any business. However, this is only a starting point. The question is does this fit in with the business’s RTO? If not, and budgets allow, could a full DR site be afforded, or could clustering be an option? Or possibly a managed standby? All of these options will help you with unplanned outages and should keep your system up and running should the worst happen.

But the reality is that in today’s IT world, planned outages result in the bulk of the downtime that can be so costly to business. Patching, updates and upgrades contribute more to database downtime than any unplanned outages. Whilst it is essential to make sure your database is running efficiently on the latest patch sets and security updates, the impact on the business of taking the database down to perform these tasks can be huge. So what can be done? A considered approach to planned downtime needs to be used. For instance, critical database maintenance can sometimes be performed whilst the instance remains online using the databases own utilities. You can also utilize your hardware’s abilities such as hardware and storage snapshots to ensure that any backup related outages could be reduced from minutes to seconds. Other things like transaction recovery whilst the database is online can also mean that the database doesn’t need to be brought down to be rebuilt from a backup or, if the budget allows, a full DR site swapped over to allowing the work to take place without impacting the business.

In summary, databases will grow in size and complexity and with this so will the risks. As DBA’s it is essential that you have the tools not only to react to these issues but also to prevent them. At RDB, we can provide solid plans and the tools required to make your essential systems stay online. For more information on this, please contact us.

Graham Barnes

Contact Us