Memory Filesystem Capacity Warning -- Troubleshooting guide

Memory Filesystem Capacity Warning -- Troubleshooting guide

Purpose

Provide a standard troubleshooting and resolution procedure when filesystem capacity on any HGX Node reaches warning levels.


When to Use

Use this KB when the monitoring system triggers the alert:

Memory Filesystem Capacity warning – For Any HGX Node

This alert indicates that one or more filesystems on the HGX Node are reaching warning capacity and may cause service disruption if not addressed.


Alert Summary

The alert is generated when the filesystem usage on the HGX Node exceeds the defined threshold.

High filesystem usage may lead to:

  • Application failures

  • Logging issues

  • Job scheduling failures

  • System instability


Severity

warning

Immediate investigation is required to prevent system disruption.


L1 Engineer Actions

When the alert is received, perform the following checks:

Step 1 – Login to HGX node

SSH into login node and then login to affected HGX Node.

Step 2 – Wait and Monitor

  1. Wait 25–30 minutes

Step 3 – Escalation (If required)

If the issue appears significant:

  • Create an internal ticket

  • Inform L2 Engineer


Step 4 – Documentation

After confirmation with L2:

  • Update the Issues Master Sheet


L2 Engineer Actions

Step 1 – Verify Filesystem and Storage Devices

Run:

df -h
lsblk
free -h

Check:

  • filesystem utilization

  • mounted storage

  • memory availability

Step 2 – Identify High Usage Filesystem

Locate the filesystem with warning usage.

Example:

du -sh /path/* | sort -hr |

Step 3 - Validate the files with customer.

Send an email to the customer including complete details of the files (such as file names, locations, sizes, and purpose), along with a screenshot for reference, before removing or clearing them. Ensure the communication is documented for KB purposes, including the reason for deletion, date/time.

Step 4 – Cleanup Space

Perform cleanup activities such as:

  • Remove old logs

  • Remove unnecessary data

  • Delete old core dumps

  • Remove unused kernel packages

Example locations:

/var/log
/tmp
/var/tmp

Step 5 – Validate After Cleanup

Verify filesystem usage again:

df -h

Ensure utilization drops below the alert threshold.

Step 6 – OEM Escalation (If required)

If usage is caused by:

  • system services

  • unknown files

  • filesystem issues

  • storage configuration problems

Escalate to OEM / platform support team.

    • Related Articles

    • Filesystem Capacity Warning in Head Node - Troubleshooting Guide

      Purpose Provide a troubleshooting and response procedure when a Filesystem Capacity Warning alert is triggered on a Head Node. This KB helps engineers quickly identify the issue, perform basic checks, and escalate when required. Alert Name Filesystem ...

    • ZFS Pool Capacity Warning Alert - Troubleshooting Guide

      Purpose This document provides a standardized approach for handling and troubleshooting the ZFS Pool Capacity Warning alert, helping L1 and L2 engineers prevent storage exhaustion and maintain system stability. Alert Name ZFS Pool Capacity Warning ...

    • Memory Filesystem Capacity Critical -- Troubleshooting guide

      Purpose Provide a standard troubleshooting and resolution procedure when filesystem capacity on any HGX Node reaches critical levels. When to Use Use this KB when the monitoring system triggers the alert: Memory Filesystem Capacity Critical – For Any ...

    • BeeGFS Capacity Warning – Troubleshooting Guide

      Purpose Provide a standard troubleshooting procedure for engineers when a BeeGFS Capacity Warning alert is triggered in the cluster. This guide ensures engineers follow consistent investigation and escalation steps. When to Use Use this KB when the ...

    • ZFS Pool Capacity Critical Alert - Troubleshooting Guide

      Purpose This document provides a standardized approach for handling and troubleshooting the ZFS Pool Capacity Critical alert, enabling rapid response to prevent storage exhaustion, performance degradation, and potential service outages. Alert Name ...