Over the last month, we’ve seen two different cases where applications were moved to a virtualization platform and in both cases, a few days after the migration, the applications had problems. The move to virtualization was immediately blamed for the problems. Coincidentally, in both these cases, our analysis using eG Enterprise proved that virtualization was not the reason for the problems.
In one case, there had been a subtle change in configuration of the servers during the migration. The change was done intentionally with the hope of enhancing the security of the systems and the administrators were unaware of the implications of the change. You can see a presentation we put together demonstrating how eG’s configuration and change tracking capability helped us diagnose the exact cause of the problem by clicking here >>.
The second case was different. Since the a new service using an application was being launched, the IT team had decided to move the application to a virtualized platform to benefit from the live migration and high availability capabilities that the virtualized platform offers. The administrators were sure that the application was working fine before the migration. With eG Enterprise in place, we were able to prove that the real issue was unrelated to the migration but that the problem was being caused by a memory leak with the application in question. The problem had not been noticed earlier but as the workload on the application increased with the new service, the problem became more severe and more noticeable.
Having the ability to look at the virtualization platform and the application in detail from the same dashboard was crucial to being able to troubleshoot both of these cases.
Our experience also highlights the challenges with virtualization. Existing infrastructures are not as well managed as they need to be – there are no baselines in place. So when problems happen after the migration, the virtualization platform is always the first to be blamed! To avoid such situations, benchmark your current infrastructure and applications for at least a few weeks before the move to virtualization.