Improved Virtual Disk Infrastructure Catalog Health Alarm due to Manual Folder Deletion
search cancel

Improved Virtual Disk Infrastructure Catalog Health Alarm due to Manual Folder Deletion

book

Article ID: 441751

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

In a VMware vSphere Foundation (VVF) environment, a persistent alarm is triggered on a block storage target, such as a Fibre Channel (FC) VMFS datastore, reading:

  • "Improved virtual disk infrastructure's catalog health alarm"

  • "Improved virtual disk infrastructure namespaces storage policy alarm"

This alert triggers after a datastore is recycled or undergoes manual administrative cleanup where the root-level /catalog or /fcd folders are removed. While the alarm is active, there is no direct impact on production virtual machine storage I/O paths, but the internal tracking database enters an unhealthful state.



Environment

  • VMware vSphere Foundation (VVF) 9.1

  • VMware ESXi 9.1.0 GA (Build 25370933)

Cause

The alarm is a genuine warning triggered when an administrator manually deletes the residual /catalog or /fcd system directories (typically via the datastore file browser) during volume recycling or cleanup operations. The vCenter Server Storage Lifecycle Manager (vslm) relies on these folders to track First Class Disk (FCD) metadata. Their removal causes the host storage daemon (hostd) to encounter an unexpected empty file state, which corrupts the FCD bookkeeping repository and forces the catalog health state to shift to unhealthy fs.

Resolution

No manual recreation of the directories or command-line reconstruction scripts are required or supported. The storage subsystem contains a built-in automated self-healing loop.

To clear the alarm and stabilize the tracking layer, execute the following steps:

  1. Allow the ESXi hosts to automatically invalidate and discard their stale hostd storage caches. This occurs naturally over time, or immediately following an administrative host daemon restart (/etc/init.d/hostd restart).

  2. Trigger any standard virtual machine provisioning task or independent disk operation. A subsequent CreateDisk infrastructure event automatically recreates clean versions of the missing /catalog and /fcd tracking directories on the root of the datastore volume.

  3. Once the folders are visible on the volume, log in to the vSphere Client.

  4. Navigate to the triggered alarm, acknowledge the alert, and manually reset the status back to green.

CRITICAL WARNING: Manual deletion of the /catalog or /fcd folders on active or recycled datastores is entirely unsupported. Do not turn off the background storage scan tracking via VsanVcMgmtConfig.xml tweaks, as this masks valid tracking anomalies.

Additional Information

Powered by Wolken Software