Deploy multi-server Lustre filesystem for large production
Budget: $15 – $20 AUD
I need an experienced Lustre engineer to take my completely bare-metal infrastructure from zero to a stable, multi-server Lustre installation that can sustain large-scale data processing workloads.
Here’s what I’m looking for:
• Architecture planning – size and lay out the metadata and object storage layers, network topology (InfiniBand or 100 GbE), and recommended OS/RAID/ZFS choices.
• Full installation and configuration of the Lustre stack across multiple servers, including LNET tuning, failover configuration, and kernel/OFED alignment.
• Benchmarking and optimisation so the file system reaches production-ready throughput and latency targets.
• A concise run-book that documents every command, configuration file, and recovery procedure so my internal team can maintain the system long-term.
I’ll provide IP ranges, rack diagrams, and remote access as soon as we kick off. Please highlight relevant production deployments you’ve completed and any performance numbers you achieved.
Here’s what I’m looking for:
• Architecture planning – size and lay out the metadata and object storage layers, network topology (InfiniBand or 100 GbE), and recommended OS/RAID/ZFS choices.
• Full installation and configuration of the Lustre stack across multiple servers, including LNET tuning, failover configuration, and kernel/OFED alignment.
• Benchmarking and optimisation so the file system reaches production-ready throughput and latency targets.
• A concise run-book that documents every command, configuration file, and recovery procedure so my internal team can maintain the system long-term.
I’ll provide IP ranges, rack diagrams, and remote access as soon as we kick off. Please highlight relevant production deployments you’ve completed and any performance numbers you achieved.