vMotion of the Instant Clone virtual machine experiences a freeze during the process.

Products

• VMware vSphere ESXi

Issue/Introduction

• Compute Live vMotion of Instant clone virtual machine goes non-responding during the migration from source ESXi host to the destination ESXi host. 
• There is a ping loss on the virtual machines during the freeze.
• After sometime the vMotion is successful and the Virtual machine resumes to work as expected.
• Content-Based Read Cache (CBRC) feature disabled on the hosts or Horizon can cause this behavior.
  
  
• Logs on the virtual machine : vmware.log

<timestamp> In(05) vmx - DISKLIB-CHAINESX : ChainESXOpenSubChain:(0) fid = 898119104, extentType = 2
<timestamp> In(05) vmx - DISKLIB-CHAINESX : ChainESXOpenSubChain:(1) fid = 971781567, extentType = 1
<timestamp> In(05) vmx - DISKLIB-CHAINESX : ChainESXOpenSubChain:(2) fid = 829437374, extentType = 1
<timestamp> In(05) vmx - DIGESTLIB-FILE : DigestLibFileOpenInt: CID mismatch -> disk=0x70c3d23b, header=0x9c423e6b.
:

:
<timestamp> In(05) vcpu-0 - Migrate: VM successfully stunned.
<timestamp> In(05) vcpu-0 - MigrateSetState: Transitioning from state MIGRATE_TO_VMX_CHECKPT (4) to MIGRATE_TO_VMX_WAIT_HANDSHAKE (5).
<timestamp> In(05) vcpu-0 - Closing all the disks of the VM.
<timestamp> In(05) worker-3880265 - Migrate: Remote Log: Destination waited for 6.02 seconds.
<timestamp> In(05) worker-3880265 - Migrate: Remote Log: Beginning checkpoint restore.
<timestamp> In(05) worker-3880265 - Migrate: Remote Log: Switching to checkpoint state. <<--There is a freeze happening around this time
<timestamp> In(05) vmx - VigorTransportProcessClientPayload: opID=mh3lvwun-1132677-auto-o9za-h5:xxxxxx-3f-24-xxx seq=879509: Receiving MKS.IssueTicket request.
<timestamp> In(05) vmx - MigrateWriteHostLog: Writing to log file took 62258 us.
<timestamp> In(05) vmx - MigrateSetStateFinished: type=1 new state=MIGRATE_TO_VMX_FINISHED
<timestamp> In(05) vmx - MigrateSetState: Transitioning from state MIGRATE_TO_VMX_WAIT_HANDSHAKE (5) to MIGRATE_TO_VMX_FINISHED (6).

Environment

vSphere 8.0

Horizon environment

Cause

The issue was caused by the CBRC (Content-Based Read Cache) filter failing to attach to the VM's disk because the cbrc_filter kernel module was not loaded on the destination ESXi host, which subsequently prevented the disk's Content ID (CID) update from being persisted in the digest file, leading to a CID mismatch.

Resolution

• Ensure the CBRC feature is enabled on the Horizon settings.Please Engage Horizon support for further investigation.
  
  
• On the ESXi host end below are the steps to enable CBRC:

Navigate to the vCenter Server UI select the ESXi host for which CBRC needs to be enabled.Validate both source and destination ESXi hosts do have this parameter set to "true".

a) Click on "Configure" -> "System" -> "Advanced System Settings"
b) In the "Advanced System Settings" menu click on "EDIT"

 For key "CBRC.Enable" provide Value as "true"

Additional Information

https://kb.omnissa.com/s/article/90715