Search datastore failed - The datastore is not completely accessible

Products

• VMware vSphere ESXi

Issue/Introduction

All VMs on one datastore are going unresponsive and when attempting to browse the datastore the following is reporting:

You will see the Host reporting the below error message:

 

Environment

VMware vSphere ESXi 7.x
VMware vSphere ESXi 8.x

Cause

The backing LUN on the storage array side is being inundated with more I/O than it is able to process.

In vCenter the datastore shows as reporting:

Lost access to volume xxxxxxxx-xxxxxxxx-xxxx-xxxxxxxxxxxx (VMFS_xxxxxxxxxxx) due to connectivity issues. Recovery attempt is in progress and outcome will be reported shortly.

In the vobd.log file the datastore reporting heartbeat loss

2026-04-27T04:15:52.513Z In(14) vobd[2097490]:  [vmfsCorrelator] us: [vob.vmfs.heartbeat.timedout] XXXXXXXXX-XXXXXXXXX-XXXX-XXXXXXXXXXXXX VMFS_xxxxxxxxxx
2026-04-27T03:23:01.482Z In(14) vobd[2097667]:  [vmfsCorrelator] 24835568598802us: [vob.vmfs.heartbeat.timedout] 6103xxxx-6dxxxx3a-7024-78xxxxxx74 nexxxxxx16_XX02
2026-04-27T03:23:01.482Z In(14) vobd[2097667]:  [vmfsCorrelator] 24835524246646us: [esx.problem.vmfs.heartbeat.timedout] 6103xxxx-6dxxxx3a-7024-78xxxxxx74 nexxxxxx16_XX02

Looking at the vmkernel.log the host reports performance issues:

2026-03-02T17:28:36.950Z In(182) vmkernel: cpu1:2097734)ScsiDeviceIO: 1799: Device naa.XXXXXXXXXXXXXXXXX performance has improved. I/O latency reduced from 345219 microseconds to 50166 microseconds.
2026-03-02T17:28:37.268Z Wa(180) vmkwarning: cpu0:2097739)WARNING: ScsiDeviceIO: 1799: Device naa.XXXXXXXXXXXXXXXXX performance has deteriorated. I/O latency increased from average value of 7300 microseconds to 312477 microseconds.

As well as reads being failed at the device layer:

2026-03-03T04:12:37.523Z In(182) vmkernel: cpu0:2097190)ScsiDeviceIO: 4595: Cmd(0x4579661e8700) 0x2a, cmdId.initiator=0x430b61e89040 CmdSN 0xda from world 10088916 to dev "naa.XXXXXXXXXXXXXXXXX" failed H:0x8 D:0x0 P:0x0 Cancelled from device layer

 

 

Resolution

Contact your storage vendor to check for issues with the LUN configuration

As a workaround, the impact on a single LUN can be reduced by migrating part of the workload to additional datastores using Storage vMotion.