High-availability for PostgreSQL in OpenShift

Running containerized applications in Kubernetes and OpenShift is a well-established way of deployment and management of software in companies, which have already adopted Could Native approach to the software life cycle. Relatively mature tools, workflows and well-known best practices are available for stateless services, such as web applications and proxies managed by automated container management platform. It is not very common to deploy stateful systems as databases into the same environment. This thesis focuses on the problematic of integrating stateful, highly available database cluster into OpenShift. Existing tools are surveyed, analyzed for their deployment in automated cloud environments and benefits as well as threats. The solution consists of container images containing selected PostgreSQL tools for high-availability and operator managing these containers in the OpenShift environment. The operator can initialize the cluster, detect crashes and autonomously perform failover.

University

Faculty of Informatics

Date of Completion

fall 2019

Resources

Leader

Vlastislav Dohnal

Consultant

Jan Horák

Student

Michal Cyprian