Cluster Handles Alerts

Prev Next

Overview

This KB covers the relevant information for the following alerts:

  1. Cluster Handles usage is above 80% /100%

    The alerts usually mean that the system is filling up and approaching the system's file handle limits.

Background

The system contains DBoxes; each DBox has a limit on handles/files, which depends on the DBox model and the SCMs' formatting layout.

Monitor

  1. For looking into the handle/inode usage, it is shown in the main dashboard in the VMS GUI - in this case, the handle/inode used space is 35.19%(relevant metric for inode usage in Analytics is Capacity,used_handles_percent)

     

    Data Reduction UI screenshot

    Data Reduction UI screenshot

  2. For tracking the handle usage of the cluster, please create a graph in Analytics, called Handle usage, which includes the following metrics: Capacity,max_handles_count, and Capacity,used_handles_count: (in this example, the max handles of the cluster are 9,129,348,300, and the used handles in the system are 3,213,059,604 → 35.19% inode usage percent.

     

    Handle count UI screenshot

    Handle count UI screenshot

Troubleshooting

  1. A new feature was introduced in 5.1 that increases the handle limit dramatically - more information regarding the limits is in the following KB - Element Limits per DBox

  2. If upgrading to 5.1 is not possible, or if your system is already on 5.1 or higher, our recommendation is to remove or delete unnecessary data or snapshots.

  3. If deleting data or removing snapshots is not an option, consider expanding the system by adding additional storage to the cluster.

Related Articles