lumi-inode-status

Check LUMI scratch and home directory inode quota usage and return a go/no-go launch verdict.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/dongzhuoyao/deepresearch --skill lumi-inode-status
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: lumi-inode-status
Source: https://github.com/dongzhuoyao/deepresearch/tree/main/.kimi/skills/lumi-inode-status
Command: npx skills add https://github.com/dongzhuoyao/deepresearch --skill lumi-inode-status

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

LUMI supercomputer file (inode) quotas are often exhausted far before block storage limits, especially for workloads that create thousands of small files like virtual environments, dataset unpacks, and large-scale generation runs. Users lack a quick, authoritative way to check their remaining inode headroom and get a clear go/no-go signal before launching these jobs, leading to avoidable failed runs and wasted compute resources.

Core Features & Use Cases

  • Dual Quota Reporting: Queries both project scratch and user home directory inode usage via native LUMI tools, plus filesystem-wide inode context for full situational awareness.
  • Actionable Verdict: Classifies quota pressure into ok/watch/warning/critical levels and returns an explicit launch safety signal, so you know immediately if you can proceed with your file-heavy workload.
  • Use Case: Before running a 100K-image generation sweep or building a conda environment on LUMI scratch, use this skill to confirm you have enough inode headroom to avoid mid-run quota failures.

Quick Start

Use the lumi-inode-status skill to check your current LUMI inode quota status and get a go/no-go verdict for your next file-heavy run.

Frequently Asked Questions about lumi-inode-status

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check my LUMI supercomputer inode quota before running a job?

To check your LUMI supercomputer inode quota, query your project scratch and user home directories to retrieve a parsed usage breakdown and an explicit go/no-go launch verdict for your file-heavy workload.

Why do my LUMI scratch jobs fail when unpacking datasets or creating virtual environments?

LUMI scratch jobs fail because inode quotas are exhausted far before block storage limits when creating thousands of small files. Verify your remaining inode headroom beforehand to prevent these mid-run quota failures.

What is the best way to avoid quota errors during large-scale file generation on LUMI?

To avoid quota errors during large-scale file generation on LUMI, perform pre-launch validation by checking your inode pressure classification to receive a clear launch safety signal before starting your workload.

Does checking inode quota on LUMI modify my system state or stored files?

Checking inode quota on LUMI does not modify your system state, as it safely queries native quota tools to report filesystem-wide inode context and usage pressure without altering stored files.

When should I verify my LUMI home directory and scratch storage file count limits?

Verify your LUMI home directory and scratch storage file count limits before any task that creates numerous small files, such as building conda environments, dataset unpacks, or large-scale generation runs.