The factor about clouds is that they can change condition even as you check out them. So it is with the Google Cloud, and the guidance we provided InfoWorld readers two many years back about producing a SQL Server failover cluster occasion (FCI) in the Google Cloud no extended reflects most effective techniques.
The guidance provided then even now operates, but improvements in the Google Cloud necessarily mean that corporations seeking a cloud-based mostly infrastructure supporting SQL Server at 99.99 per cent availability ranges now have superior, far more effective methods to produce SQL Server FCIs. The fundamentals of these new tips can be identified in Google’s on line guidelines for configuring a SQL Server FCI.
Google’s existing guidance is very clear and effortless to adhere to. But corporations really serious about 99.99 per cent ranges of availability should really adhere to these tips only up to a place. As outlined in the documentation, the FCI depends on shared storage configured utilizing Microsoft’s Storage Areas Direct option, and that results in an availability Capture-22.
The concern is this: Microsoft tips for Storage Areas Direct involve all the VMs and storage to reside in the similar facts middle, and it is not very clear that an FCI configured in the method outlined by Google complies with this necessity. The Google documentation directs you to put the VMs in various zones inside the Google Cloud, and if these zones are not the similar facts middle then you’re producing a Storage Areas Direct configuration that is unsupported.
The Capture-22 is that if you configure the FCI so that all the VMs are in the similar facts middle, and therefore compliant with Storage Areas Direct tips, your FCI does not qualify for Google’s large availability SLA (which involves that at least 1 VM in the SQL Server FCI reside in a distinct zone). If the full facts middle goes offline (and full facts centers have completed so), the entire FCI will also go offline.