Linux infrastructure and research operations
I work alongside the people keeping research systems usable: accounts, software, node images, monitoring and the operational details that make a cluster dependable.
Help with the work between deployments
A cluster’s daily usefulness depends on more than its installation. Researchers need software that loads correctly, jobs that reach their data, accounts with the right access and someone who can investigate a failure across those boundaries.
I work on Linux and UNIX systems, identity integration, software environments, automation and monitoring. My experience includes Ansible, Puppet, Prometheus and Grafana, as well as node-image and provisioning workflows. We agree which systems and tasks are included before work begins.
Investigate a problem that crosses layers
Some faults appear to belong to an application but originate in its environment: a module points to a removed library, a batch shell loads different settings from a login shell, or an image rebuild erases a local repair. I trace the running path and test the change where the researcher will use it.
For scheduled automation, I check what ran and what it produced, including reporting and recovery after interruption. For management tasks, the scope can include read-only inventory or agreed Redfish workflows, with the target and maintenance requirements established before disruptive work.
Campus and inter-site networking
Research traffic also travels between instruments, offices, compute systems and collaborating sites. I design and troubleshoot those connections as well as the network inside a cluster, using Juniper, Arista, Cisco and Brocade platforms where appropriate to the existing environment.
We agree which path and applications are in scope, then follow the traffic to the point where access or performance breaks down. Campus and inter-site work is scoped separately from the fixed compute-fabric diagnostic; the qualification conversation helps identify which engagement fits.
Leave changes that survive maintenance
A live repair needs a home in the configuration or provisioning source that will recreate it. I help teams connect the running system with images, automation and documentation, and verify behavior after the relevant lifecycle change.
Monitoring follows the same approach. A service being up does not show that a scheduled process is fresh or that an application can complete its work. Checks should help the next operator locate the problem and understand what requires attention.
Start with a manageable scope
A focused day or week can be a useful starting point for one operational problem. Larger migrations, ongoing requirements or infrastructure build-out are discussed and quoted for their actual scope. I also help with administrator handover, vendor discussions and planning around the team’s time and budget.
The ordering page describes the available engagement formats and current terms. If the issue spans storage, fabric or the scheduler, we can select the appropriate diagnostic rather than stretching a small engagement beyond what it can reasonably cover.
Engagement options
Choose a format below to see current rates and the ordering options for that engagement. If you are unsure which fits, we can work through the scope together.
- Focused engagement — one dayA bounded problem or second opinion, worked through over one day.
- Focused engagement — one weekA focused week for investigation or implementation with an agreed scope.
- Cluster design and build-out — any scaleArchitecture and delivery scoped and quoted for your system.