Database administrators [DBAs] across the world have shared the best ways to ensure that database virtualisation is as simple of possible with a quick fire list of five different factors to consider.
SolarWinds took a survey of 498 IT professionals that found 76 per cent of DBAs have already virtualised databases or are utilising a hybrid approach of virtual and physical databases to support their environment.
The list of five best practices that DBAs think should be adopted by all those handling a database virtualisation project for it to go smoothly are as follows:
1. Don’t fear virtualising a large database with a high transactional load
The current virtualisation technology available is both reliable and robust, and brings performance that compares very closely to that enjoyed when running a database on a virtual server.
2. Develop a good working relationship with the virtualisation administrator
If the DBA can’t make changes easily to the virtual machine resources then the biggest benefit of virtualisation is immediately lost. Unlike a physical server additional CPU, RAM and storage resources can be added to a virtualised database server.
3. Utilise shared metrics to improve cross-collaboration between IT groups
Establish a unified understanding among the database team, developers, system and virtualisation administrators to make sure there are no disagreements or confusion. Teams can work together in order to stay on top of various key considerations including server resource contention and I/O bottlenecks.
4. Correlate physical host and VM resource metrics with SQL query performance
VM metrics are sometimes misleading due to the OS not accurately reporting resource metrics when running on a VM. By putting metrics in the context of query performance the impact of various issues caused by noisy neighbours, VM events, and resource provisioning can be identified quickly.
5. Monitor SQL response time before and after virtualisation
Porthole AdIt’s imperative to use the correct tools to gain visibility on response time. This information gives evidence to suggest whether database performance is being maintained or improved. It also makes it easy to pinpoint errors and resolve issues when performance issues are occurring.