Problem description
VMware virtual machine backups intermittently fail to use the preferred HOTADD transport mode, falling back to the slower NBDSSL method. This troubleshooting guide addresses one of the most common causes for this issue, specifically when the error "Cannot access datastore" is observed in the logs.
This issue may affect some VMs while others on the same ESXi host back up successfully using HOTADD. The problem is particularly noted on VMs with virtual disks spread across multiple datastores.
Cause
While there are several potential reasons for HOTADD to fail, the specific cause addressed in this article is twofold:
Incomplete Datastore Presentation: The ESXi host on which the backup proxy VM is running does not have access to all datastores that the target virtual machines are using. For HOTADD to function, the proxy's host must have a direct connection to and be able to mount all datastores where a VM's disks reside.
"All-or-Nothing" Transport Mode Selection: For a single VM backup job, if even one of its virtual disks is on a datastore that is inaccessible for HOTADD, the backup software downgrades the transport mode to NBDSSL for all disks of that VM. This happens even if other disks for the same VM are on perfectly accessible datastores.
Traceback:
This specific issue can be confirmed through the following log files on the backup proxy.
vixDiskLib logs will show the primary error indicating the datastore is not accessible, leading to the fallback.
Log file: vixDiskLib-*.log
Snippet:
VixDiskLibVim: SupportHotAddTransport: Cannot access datastore for one of the disks of VMDK '[DATASTORE_NAME] VM_NAME/VM_NAME-000002.vmdk'.VixDiskLib: Falling back to NBDSSL transport.
EnterpriseWorkloads logs confirm the final transport mode used for each disk during the job.
Log file: EnterpriseWorkloadsFL.*.log
Snippet:
level=info ts=2025-08-20T18:10:17.1104646Z filename=file.go:114 message="Transport Mode used for disk [DATASTORE_NAME] VM_NAME/VM_NAME-000002.vmdk: NBDSSL"
Resolution
To resolve the datastore accessibility issue, you must ensure that all datastores containing disks for the virtual machines being backed up are presented and mounted to all ESXi hosts where the backup proxy may be running.
Identify the Backup Proxy Host: In the vSphere Client, locate the backup proxy VM and identify which ESXi host it is currently running on.
Identify the Inaccessible Datastore: For a failing VM, list all datastores its virtual disks reside on.
Verify Datastore Mounting: For each datastore identified in step 2, select it in the vSphere Client, navigate to the Hosts tab, and verify that the backup proxy's host (from step 1) is listed.
Mount Datastore on Host: If the host is missing, right-click the datastore and select the option to Mount Datastore to Additional Hosts. Select the required ESXi host and complete the wizard.
Rescan Storage: On the ESXi host where you just mounted the datastore, perform a rescan of all storage adapters to ensure the change is recognized immediately.
Re-run Backup: Trigger the backup job again. It should now successfully use the HOTADD transport mode.
Note: Before performing storage operations in your vSphere environment, it is highly recommended to consult official VMware documentation and your organization's guidelines to prevent unintended impact.