Cannot see some or all storage devices in VMware vCenter Server
search cancel

Cannot see some or all storage devices in VMware vCenter Server

book

Article ID: 344351

calendar_today

Updated On:

Products

VMware vCenter Server VMware vSphere ESXi

Issue/Introduction

You cannot see some or all storage devices or the list is incomplete and missing the intended device(s) when you try to perform these operations:
  • Creating or formatting a new VMFS datastore using the Add Storage feature.
  • Extending an existing VMFS datastore by using another distinct LUN as an extent.
  • Extending an existing VMFS datastore by utilizing unallocated space made available within the same LUN (via a previous enlargement process via the SAN array).
  • Creating Raw Device Mappings (RDMs) for use with Virtual Machines.
  • In ESXi, the device can be seen by the ESXi host, but the datastore is inactive under storage.
  • You may an error similar to this in the vmkernel.log:

    cpu0:946544)ScsiDevice: 4592: naa.60a9800037542d68492############# device :Open count > 0, cannot be brought online
You have already:
  1. Presented one or more LUNs on the SAN array and presented it to your VMware ESXi Servers.
  2. Connected your VMware vSphere Client to the VMware vCenter Server hostname or IP address.
  3. The VMware ESXi servers have rescanned their HBAs from the Configuration tab's Storage Adapters view.
  4. The LUNs are visible under Storage Adapters.

Environment

  • VMware vCenter Server 6.x
  • VMware vCenter Server 7.x
  • VMware vCenter Server 8.x
  • VMware vSphere ESXi 6.x
  • VMware vSphere ESXi 7.x
  • VMware vSphere ESXi 8.x

Cause

The general cause of this issue is the non-uniform presentation characteristics of the SAN LUNs across the managed VMware ESXi servers. A requirement is having a uniform or consistent presentation of LUNs across all managed VMware ESXi Servers. Specifically, the LUN ID for each respective LUN must match between all VMware ESXi Servers.

Identifier Specifics:

VMware ESXi above rely on the Understanding SCSI Check Conditions in VMkernel logs during rescan operations device identifier to identify and distinguish between multiple devices. Supported identifiers are:
  • T11 Network Address Authority (NASA and NAA) 64 and 128-bit integer values, such as:
    • naa.60060160c5602100ce############## or vml.020001000060060160c5602100ce##############5241494#####
  • Extended Unique Identifier (EUI) such as:
    • eui.02004567A#######
  • Storage Name Service (SNS) such as:
    • iqn.2001-05.com.vmware:storage.san.sys.xyz
  • T10 identifiers.
  • iSCSI Qualified Name (IQN) identifiers.
If the device does not support VPD page 0x83, then it will be assigned an MPX identifier value, such as
mpx.vmhba0:C0:T0:L0.
This is typically the case for local disks which could not be physically shared, rather than SAN LUNs.

Resolution

Background

VMware vCenter facilitates numerous storage filters to determine using devices for potentially destructive operations such as formatting, creating extents, and creating Raw Device Mappings (RDMs) for use with VMs. There are additional filters in place to account for presentation inconsistencies (or non-uniform presentations) among managed VMware ESXi hosts in the datacenter.
 
Rules and considerations for these filters are:
 
VMFS Filter config.vpxd.filter.vmfsFilter
  • The LUN is already being used as a datastore on the same host.
  • The LUN is already being used as a datastore on a different host.
RDM Filter config.vpxd.filter.rdmFilter
  • The LUN is already being used as an RDM on the same host.
  • The LUN is already being used as an RDM on a different host.
Rescan Filter config.vpxd.filter.hostRescanFilter
VMware ESXi performs automated rescans of VMFS datastores after any datastore configuration operation, including Create, Expand, Extend, Remove, and when upgrading to newer versions of VMFS.
This helps vCenter to offer centralized management by ensuring that ESX hosts provide consistent views of the storage available to them.
Same Host and Transports Filter  
The operation of extending a VMFS datastore can utilize entire Logical Unit Numbers (LUNs) and partitions. This filter ultimately prevents corruption and degraded performance. There are several cases where this filter is triggered:
  • Multiple VMware ESXi hosts managed by VMware vCenter can see a datastore but the extent candidate is visible only on some of the hosts, rather than all.
  • Your datastore is a local transport type (such as Block or Parallel) and you are attempting to extend it with a datastore of a shareable transport type (such as Fibre Channel or iSCSI), or vice-versa.
  • There is no uniform presentation between multiple ESXi Servers.
In effect, this filter will remove extent candidate LUNs which are not seen on all of the same hosts as the datastore that is being extended. The non-uniform presentation will result in this filter taking action and declining candidacy for several of the operations, either making the candidate list partial, or empty.

For more information, see Storage Filtering.

If your host LUN IDs are not consistent across the managed VMware ESXi Servers, schedule downtime and configure them for consistent or uniform presentation, ensuring each managed ESXi hosts see each respective LUN with the same LUN ID.
If necessary, work with your storage array vendor to perform this operation.

To verify whether presentation efforts are resulting in inconsistency across the datacenter, see: Identifying disks when working with VMware ESXi.
 
For further information on presentation specifics, see the vSphere Storage for vSphere and Fiber Channel SAN Configuration Guide or the iSCSI SAN Configuration Guide for vSphere.
Powered by Wolken Software