Keep VMware tools, as well as select hardware and software, safe and organized with a management cluster. If an issue comes up, use it to bring systems back online.
Using a VMware management cluster is a good way to organize management tools and separate important software and hardware in the event of an issue.
The growth of the virtual environment has spurred growth in the number of tools used to manage these environments. Management tools add to the already large pools of servers, management services and infrastructure that a growing business needs to function. A VMware management cluster can enable administrators to keep track of these tools -- even in the event of system failure.
The utility servers that modern data centers need experience sprawl just like application servers do. Management and infrastructure servers are critical to the data center, but that doesn't mean they are heavily used. Often, these servers need to be there even though their workloads are light. The modern virtualized data center has high availability (HA), failover hardware, storage controllers and networking, so large-scale outages are unlikely.
These tools and features -- for both management and recovery -- present challenges. Placing management tools in the same environment they manage is like locking the keys to a house inside that house. The argument for relying on technology such as HA restarting critical VMs is valid if the infrastructure is working and there aren't hundreds of VMs tagged as high restart. Problems can arise, however, if it's a subsystem, such as primary storage, that has issues.
Placing management tools in the same environment they manage is like locking the keys to a house inside that house.
A functional management console could resolve it, but the VM might be offline because the hosts booted up before the primary storage came online. Admins can perform another vCenter scan to fix this, but only if vCenter is working. Without management tools, the challenge admins face is even more daunting.
Use a VMware management cluster to prepare for issues
An effective VMware management cluster assumes that something critical will fail and that preparation is necessary. As more of the data center is virtualized, events demand more foresight.
There are a few different perspectives on using a VMware management cluster. One strategy involves a complete transportation of existing management tools to a new environment. Another option is a hybrid approach with redundancy and shared servers and duties. The decision hinges on the degree of separation the admin wants.
Besides the separation of roles, there's also physical separation, such that each environment has its own storage, networking and compute that are fully redundant to environments with key, shared subsystems.
Completely separating the VMware management cluster from the production environment offers the best protection. Admins can take a few shortcuts to achieve this. For example, rather than having shared storage, admins can use VMware vSAN or other local storage options. A VMware management cluster usually contains tools and servers that aren't as workload-heavy, so admins can bend some rules for performance or failover in favor of cost savings and reduced complexity. Much of the decision depends on costs and available hardware.
What's included in the VMware management cluster is up to each organization, but a linked vCenter should be top priority. Consider infrastructure, as well as critical tool servers, for networking and storage management. An Active Directory controller, along with backup domain name system and Dynamic Host Configuration Protocol servers, would keep systems online and accessible. Don't forget less common servers, such as emergency print servers or communication servers. The list doesn't have to include everything in the production environment, but it should have enough to maintain the system while admins work to bring everything back online.
A VMware management cluster might contain six or more virtual servers -- only a fraction of what's necessary to manage a data center -- but keep the purpose of it in mind. The management cluster should offer the tools necessary to bring the core production systems back online. It's a set of tools that enables repair, not running the entire data center long term. Ensure IT staff and company leaders understand these differences to alleviate confusion about what can and can't be done before an issue occurs.