Experiencing Corruption After Taking Snapshots Across Multiple Sites/Hosts/VMs

Products

• VMware vSphere ESXi

Issue/Introduction

In case of a VM with VDISKS residing on 4KN SCSI RAID LUN (CISCO RAID CONTROLLER), after taking a VM snapshot, Sesparse reports corruption. This is seen on multiple hosts.

Hostd.log

 

2024-11-15T18:42:44.008Z verbose hostd[2100752] [Originator@1176 sub=Vmsvc.vm:/vmfs/volumes/#######-####-####-####-############/datastores reported issue are/test.vmx opID=########-#######-auto-1d3nr-h5:10222459-af-01-1-abcd] Handling vmx message 3262: The redo log of 'test-01-000001.vmdk' is corrupted. If the problem persists, discard the redo log.

 

Vobd.log

 

2024-11-15T18:53:39.746Z: [UserWorldCorrelator] 111111973677us: [vob.uw.core.dumped] /bin/vmx(2255035) /vmfs/volumes/#######-####-####-####-############/test/vmx-zdump.001

Environment

ESXi 7.x

ESXi 8.x

Cause

The corruption is in a 512-byte sector in the Sesparse metadata region. When this pre-zeroed 512-byte sector in the delta was read from the disk the data returned was junk. An immediate re-read returns expected zeroed block indicating on-disk data is good.

Issue  observed with a datastore created  with CISCO RAID controller .

 

Resolution

 suggested to contact CISCO hardware support for further troubleshooting