Ask the Expert - Mr. Database
Automation: New and Improved
by Eric Gross, DBA

The new features in the most recent release of GridApp Clarity™ simplify the day-to-day management of databases. Tasks such as creating a multi-node Microsoft SQL Server clustered instance (which used to take days of following a run book) now is refined into a single XML template that can be used to perform the activity time after time in the same exact way. We refined all of the available inputs, such as vendor documentation and practical testing into Clarity to enable the enterprise to push out procedures more reliably then ever before. Clustered processes can have thousands of discrete steps, so it is important to check that there is a good likelihood of success before entering into a process. In this release, Clarity has the capability to check for a large number of requirements before a job even starts, giving the user the opportunity to resolve the problems. Otherwise, if the user is confident that the problem will be resolved by a custom script that is run before the remainder of the activities, then the issue can be ignored – it will be checked again before it is needed.

Based solely on the huge number of inputs and steps, it isn’t practical to perform all these activities manually – each operator would have a slightly different interpretation of the steps resulting in inconsistent deployments. Run books have a use – they are a reference for creating automation of the included steps. The most time-consuming portion of automation is gathering all of the requirements and subsequently working through the various steps that get you from here to there.

Other automation alternatives, such as Grid Control (by and exclusively for Oracle software) attempt to satisfy the need to perform repetitive tasks, and to some degree they succeed. The fundamental difference between other solutions and Clarity is that the intelligence built into Clarity enables the user to perform a wide swath of the goals without writing a single line of code. Rather than exposing a framework to developers who need to write code to do what is required, Clarity intrinsically has the functionality to implement the requirements.

Generalized automation solutions have limitations that interfere with complex applications in different ways. Most solutions allow only a single type of entity per host. This is insufficient when you need multiple instances on the same host and you don’t want to model each database instance as a different entity. In other cases (clusters for example), entities need to be able to span multiple hosts – and we’re not just talking about showing the same entities on one host, but adding the functionality required to add and remove the elements on arbitrary nodes while leaving other hosts unchanged. The basic model is zero or one entity per host – it is so simplistic that a specialized automation application is required to properly manage the database layer. Since these solutions can’t understand complex application structures there is obviously no automation content to enable advanced automation. Far from it, they don’t even have a framework for content to be created for enabling these types of uses which are becoming more common as datasets and uptime requirements become more extreme.

<< back