Australian cloud for business software — built and run in Sydney since 2002

NetLeverage
NetLeverage

vSphere HA: number of heartbeat datastores for this host is less than required: 2

Technical article
VMware

About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.

A host in a vSphere HA cluster shows The number of vSphere HA heartbeat datastores for this host is 0, which is less than required: 2 (the wording in Broadcom KB 392502). KB 318871 prints The number of heartbeat datastores for host is 1, which is less than required: 2. The count varies, and "required: 2" is the default. Broadcom says vSphere HA needs at least two datastores that all hosts in the cluster can reach. Add or select a second shared datastore. An advanced option hides the message, and the missing second heartbeat path stays.

What the message means

When a primary host cannot reach a secondary host over the management network, vSphere HA needs a second way to tell what happened. Broadcom's vSphere 8.0 guide says the primary host "uses datastore heartbeating to determine whether the secondary host has failed, is in a network partition, or is network isolated." If the secondary has stopped datastore heartbeating, "it is considered to have failed and its virtual machines are restarted elsewhere."

vCenter Server selects a preferred set of datastores for each host. By default the count is two, and the advanced option das.heartbeatdsperhost changes it. Broadcom's HA advanced options page gives the valid range as 2 to 5 and the default as 2. KB 313261 adds that HA reports a host config issue "if it was not able to select the required number of datastores for a host given by das.heartbeatDsPerHost".

Broadcom's KB gives the rule: "As of vSphere 6.0 and above, vSphere HA requires a minimum of two shared datastores accessible by all hosts within the cluster to handle datastore heartbeat monitoring." KB 318871 adds that the message "can also occur if one of the two heartbeat datastores go offline due to underlying storage issues."

Two limits from the 8.0 guide matter here:

  • "A vSAN datastore cannot be used for datastore heartbeating." If the only shared storage is vSAN, there can be no heartbeat datastores. Storage that is reachable by a network path independent of the vSAN network can be used instead.
  • vSphere HA creates a directory named .vSphere-HA at the root of each datastore it uses. The guide says "Do not delete or modify the files stored in this directory".

NetLeverage runs VMware vCenter, vSphere and ESXi.

Likely causes

Cause What you see How to confirm
Fewer than two datastores are shared by all hosts Count 0 or 1 on every host The Heartbeat Datastores list shows fewer than two datastores, or the hosts pane shows datastores not reachable by every host
The selection policy is limited to a list with one datastore Count 1 although other shared datastores exist The cluster's Heartbeat Datastores option is set to use datastores only from a specified list
One of two heartbeat datastores went offline Warning appears on a cluster that was clean The datastore shows as inaccessible on one or more hosts
The only shared storage is vSAN Count 0 The cluster has no datastore other than vSAN visible to all hosts
das.heartbeatDsPerHost was raised above the shared datastores available Count 2 against a required 3, 4 or 5 The cluster's Advanced Options list contains the option

Prove the cause before changing anything

All of these are read-only. Do not click OK on any edit dialog while looking.

  1. In the vSphere Client open the cluster, then Configure, then vSphere Availability. KB 392502 says to expand "Datastore for Heartbeating" and check whether any datastores are selected.
  2. Click Edit for vSphere HA and go to Heartbeat Datastores. Read the selection option and the Available heartbeat datastores list. Broadcom's configuration page says "When a datastore is selected, the lower pane displays all the hosts in the vSphere HA cluster that can access it." A usable datastore lists every host. Close the dialog without saving.
  3. In the same Edit dialog, read Advanced Options. Note whether das.heartbeatDsPerHost or das.ignoreInsufficientHbDatastore is present.
  4. On the host named in the message, list its datastores and backing devices. KB 316540 gives this command and says it lists "extents for each volume and mapping from device name to UUID":
esxcli storage vmfs extent list

A good result is at least two datastores that appear on every host in the cluster. A bad result is one, or datastores that only some hosts see. A datastore on a device local to one host does not count. If a datastore that should be shared is missing from a host, treat it as a storage presentation or access fault first. The guide "APD or PDL: classify an inaccessible ESXi datastore" covers how to classify that.

Fix

Choose the fix that matches the cause. The first two remove the cause. The third only hides the message.

Present a second shared datastore

The remedy that removes the cause is a second datastore that every host in the cluster can see, presented through the storage array or NFS server. That storage change is outside this guide. Once it is in place, you can run esxcli storage vmfs extent list on each host again.

Change the heartbeat datastore selection

KB 318871 offers two selection options: "Automatically select datastores accessible from the hosts" or "Use datastores from the specified list and complement automatically if needed". KB 318871 says that with the second option at least two datastores must be selected from the list below it.

First save the current setting. In the Heartbeat Datastores page, write down the selected option and every datastore ticked in the Available heartbeat datastores list.

Risk: the next steps change which datastores vSphere HA writes heartbeat files to for the whole cluster. A datastore that not every host can reach cannot serve as a heartbeat datastore for the hosts that cannot reach it.

In the vSphere Client choose the cluster, Configure, vSphere Availability, Edit for vSphere HA, Heartbeat Datastores. Pick one of the two options above. If you pick the specified list, tick at least two datastores that list every host. Click OK.

To undo it: open the same page, set the option and the ticked datastores back to what you wrote down, and click OK.

Hide the warning (does not add a heartbeat path)

KB 318871 describes this under "To only disable the HA error message". The 8.0 advanced options page describes das.ignoreinsufficienthbdatastore as one that "Deactivates configuration issues created if the host does not have sufficient heartbeat datastores for vSphere HA. Default value is false." KB 313261 says the option needs HA reconfigured on all hosts in the cluster.

First save the current state. Write down every option and value in the cluster's Advanced Options list, and note whether vSphere HA is turned on.

Disruptive step: the next steps turn vSphere HA off for the whole cluster and back on. While it is off, HA does not do what it does after a host failure, which is to restart the virtual machines elsewhere. Every host unconfigures and then reconfigures its HA agent. Do this in a quiet window, with no host or storage maintenance in progress.

Follow KB 318871: in vSphere Availability click Edit for vSphere HA, open Advanced Options, type das.ignoreInsufficientHbDatastore as the option and true as the value, click Add and OK. Click Edit again, switch vSphere HA off, click OK and allow time for all hosts to unconfigure. Then click Edit, switch vSphere HA on, click OK and wait for all hosts to configure. Check that vSphere HA shows as turned on.

To undo it: reset the option to its default value in the vSphere Client, so that the list matches what you wrote down. Then reconfigure HA on all hosts in the cluster (right-click each host in the vSphere Client and choose Reconfigure HA), which KB 313261 lists as the requirement for this option, and check that vSphere HA is on.

Example (illustrative): the cluster cl-prod has esx01, esx02 and esx03. Each host has one local datastore (ds-local-esx01, ds-local-esx02, ds-local-esx03) and all three see one shared datastore, ds-shared01. Each host shows the message with a count of 1. The default required count is 2, so 2 - 1 = 1 more shared datastore is needed. The storage administrator presents ds-shared02 to all three hosts. In Heartbeat Datastores the administrator chooses "Use datastores from the specified list and complement automatically if needed" and ticks ds-shared01 and ds-shared02, with all three hosts listed against each. The count becomes 2 and the message clears.

Confirm the fix

  1. The host summary no longer shows the message. If you hid it with the option, this proves nothing about heartbeat paths.
  2. The Heartbeat Datastores page lists at least two datastores, and the hosts pane for each lists every host.
  3. esxcli storage vmfs extent list on each host shows both datastores.
  4. vSphere HA shows as turned on and each host as configured.

What to watch afterwards

  • Storage maintenance on one of the two datastores brings the message back. Check the count before and after it.
  • If someone raises das.heartbeatDsPerHost, the required count rises with it, so add datastores to match.
  • If you used the advanced option, record it. Anyone who later reads a clean host summary will not know a heartbeat path is missing.

Risks and limits

  • These commands have not been run or lab-tested.
  • Broadcom's KB 318871 and KB 392502 were both updated 10-08-2026. KB 318871 names vCenter Server 6.5.x to 8.x, and KB 392502 names ESXi 7.x to 9.x.
  • The 8.0 pages give the option as das.heartbeatdsperhost and KB 313261 as das.heartbeatDsPerHost. KB 313261 lists that option for versions up to 6.0, while the 8.0 page lists it without a version range. Use the spelling the Advanced Options dialog in your vCenter Server accepts.
  • Presenting storage to hosts is not covered. The steps for that depend on the array or NFS server.

Sources