Title here
Summary here
Developer Blog
Code Deep Dive: Dynamic Data Persistence for our Control Plane
This blog post dives into the 'In-Memory Working Set with Periodic Snapshots' pattern implemented in the node-orchestrator project. Learn how this approach leverages DuckDB and Parquet to efficiently manage control plane data, offering speed and reliability. Discover the advantages for HPC sysadmins, including reduced management overhead and simplified recovery, as well as comparisons to other data storage methods.
Service-to-Service Auth in OpenCHAMI: Client Credentials, Simple and Safe
A short guide to use OAuth2 client credentials to call another service with a JWT in OpenCHAMI.
Turnkey OpenCHAMI: A Weekend Deploy for a Small HPC Cluster
A practical, short path to try OpenCHAMI on a lab cluster or sandbox.
Exploring Multi-tenacy capabilities in SMD with Groups and Partitions
SMD comes with a built-in partitioning feature that should allow us to partition off certain components through it's API. In this blog post, we explore how groups and partitions work in practice.
Boot Script Service: Switching from Etcd to PostgreSQL for a Smaller Storage Backend
As part of LANL's exploration of the CSM codebase, they replaced Etcd with PostgreSQL for BSS persistence and learned a lot about software boundaries along the way.
Exploring CSM Microservices for OpenCHAMI
At LANL, we used two microservices from HPE's CSM to do discovery and boot of a ten node cluster and we learned so much along the way about redfish, postgres, httpboot, aws, and of course, ourselves.
Magellan: Alternative Redfish Discovery for OpenCHAMI
As part of LANL's exploration of the CSM codebase, they found it necessary to augment the built-in discovery features of CSM with a new standalone command based on gofish and bmclib.