GP - Fotolia
Industry buzz around containers and Kubernetes can make it difficult for IT professionals to uncover best practices to use these technologies. This is especially true in the critical area of backup and recovery.
Containers and Kubernetes, frequently used interchangeably in the context of storage and data protection, are not the same technology. Kubernetes is a tool to manage and orchestrate the execution of containers. Containers require data protection, including backup, because, while their data has been predominantly ephemeral, persistent storage for containers has come into focus.
It can be challenging for admins to establish the need for container data protection as part of an overall backup and recovery strategy. At many organizations, data protection is an afterthought for new workloads, especially since, traditionally, containers have not used persistent storage. In addition, Kubernetes clusters are architected for high availability. Some IT professionals, as a result, assume container environments do not require backups.
This couldn't be further from the truth. Container-based workloads are beginning to generate mission-critical data and require persistent storage. This data must be recoverable in the face of disruptive events, including infrastructure outages, cyberattacks and accidental or malicious data deletion.
Container data protection challenges and tips
Container environments operate at massive scale in terms of the number of container instances that comprise various application components. Rather than an entire app being mapped to a set of VMs or to a physical system, various app components are distributed across multiple systems for fault tolerance and load balancing. This creates challenges for hypervisor- or system-centric backup tools, especially those that require an agent on each server. Admins can also create and destroy containers instantaneously and at will, resulting in a dynamic environment that changes frequently.
To establish a container data protection strategy, IT professionals must protect everything the application or database needs to run. These resources lie both inside and outside of the container cluster in external storage and databases. It is difficult to achieve consistent backups for applications or databases due to the rate at which data changes. Also, it is not feasible to install an agent to execute an application-consistent backup on each container.
Always test for backup integrity and recoverability. Container data protection should be part of the overall backup and recovery implementation -- from the beginning. This provides confidence in the ability to recover, as well as clarity in the recovery process.
A variety of vendors provide container backup capabilities, resulting in a crowded market. Container data protection vendors including Cohesity, Commvault, Dell Technologies, Druva, HYCU, IBM, Veeam, Veritas and Zerto are adding Kubernetes support. Newer vendors specializing in Kubernetes backup include Trilio; Kasten, which was acquired by Veeam in 2020; and Portworx, which was acquired by Pure Storage in the same year.
To choose a vendor, first consider how the container data protection tool will fit into the organization's broader backup and recovery implementation. In addition, ensure the product can meet required service levels for recovery point objectives and recovery time objectives.