IT admins and business owners are constantly thinking about how to make business services and applications more highly available to their end users, and there are a ton of different disaster recovery and high availability solutions out there in the marketplace today. Across all these different solutions, each approach comes with its pros and cons and different management struggles. Clustering is probably the most widely known availability technology but replication solutions have come on strong the past several years. When implementing any solution, you have to make sure that these don’t impact standard maintenance procedures in production IT environments. This primarily consists of the Windows Update patches every month, and sometimes even maintenance to the server hardware.
With solutions like traditional failover clustering, you have to patch all nodes of the cluster as the operating system is in its own volume but the data is shared. Because there is more than one machine capable of holding the “active” role, admins are able to patch servers that are in “passive” roles and then make them active after patching is complete. By leveraging the multi-node active/passive cluster, admins are able to minimize downtime and mitigate risk of bad patches.
However, in today’s world of virtual machines, many DR solutions work by replicating simply VM images from one host to another. With these replication solutions, the monthly patching cycle can be more complicated depending on what is being replicated. VMware vSphere replication replicates the entire VMDK, or virtual machine image, so you only need to patch the active copy of the server. The pitfall here is that if the patch is bad or it breaks an application there isn’t any easy way back as the problem has then be replicated to the other copy of the VMDK. The admin will either need to fix the problem in place or go back to earlier copy of the image and try again. Either way, there is potential for extended downtime or even data loss.
Host based replication solutions that focus just on the data of the application are a hybrid of both these solutions. These types of solutions, of which the Neverfail IT Continuity Engine is an example, combine the concept of active/passive clustering and replication of shared data between the cluster nodes. One advantage of these solutions, besides protection from application downtime, is the ability to test patches and have a second copy of the data somewhere else. One challenge faced by this architecture is the task of patching passive servers that may can be hidden from the network and thus unable to contact the Windows Updates server for patches.
Neverfail has provided a solution to this standby or passive server maintenance challenge with their new patching utility for the Heartbeat / IT Continuity Engine product. With the utility in place on a Neverfail clustered pair, the IT admin gains the ability to have the patching ease of clustering with the functionality of a host based replication solution. In this scenario, the monthly security updates to the operating system can be applied and tested on the passive servers before being rolled out in production. In addition, by leveraging the Neverfail pair as part of the patching process, the planned downtime and uncertainty of the patch/reboot cycle can be avoided on the active servers. Just one more way that Neverfail IT Continuity Engine protects businesses from downtime and keeps users connected to their applications regardless of the threats to availability!
