Portworx has announced the release of STORK, its new open-source Storage Orchestration Runtime for Kubernetes solution.
STORK was developed in collaboration with customers running large-scale stateful applications in production who were looking to address operational issues.
“We listened carefully to our enterprise customers and built STORK to solve critical efficiency limitations that enterprises running stateful applications on Kubernetes commonly encounter,” said Gou Rao, CTO of Portworx. “We’re committed to expanding on this open source project with the help of the community in order to address the pain points and needs of enterprises.”
The solution uses Kubernetes to enable DevOps teams to run stateful applications more efficiently with hyperconvergence, failure-domain awareness, storage health monitoring and snapshot-lifecycle features.
In addition, it uses a Kubernetes scheduler extender to influence pod scheduling so that it can ensure hyperconvergence during failover, and manage anti-affinity to ensure pods are scheduled across failure domains.
According to the company, a common problem with stateful apps is that the overall health of a pod will wear and tear over time. To address this, STORK fails pods when the storage driver enters an error or unavailable state, which allows applications to be highly available without user intervention.
Other features include support for volume snapshots through Kubernetes in order to automate complex data workflows, and the ability to communicate with storage drivers through a plugin interface to allow a choice of drivers beyond Portworx.