Database VM reporting in guest I/O errors
Article ID: 394823
Updated On:
Products
Issue/Introduction
- VM resides on vSAN datastore
- read write issue within the guest OS
- Object Storage Policy is RAID0
- vSAN Skyline Health is reporting Physical disk issues and Object Health alerts
- vmmkernel logs shows the failed disk suffered a medium hardware failure
2025-04-19T16:42:45.526Z cpu4:2106057 opID=bc9005a)Partition: 433: Failed read for "naa.################": I/O error
2025-04-19T16:42:45.526Z cpu4:2106057 opID=bc9005a)Partition: 1109: Failed to read protective mbr on "naa.################" : I/O error
2025-04-19T16:42:45.526Z cpu4:2106057 opID=bc9005a)WARNING: Partition: 1262: Partition table read from device naa.################ failed: I/O error
2025-04-19T16:42:45.526Z cpu47:2099121)ScsiDeviceIO: 4062: Cmd(0x45a1d8a06580) 0x28, CmdSN 0x1 from world 2106019 to dev "naa.################" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x3 0x31 0x0.
- 0x3 0x31 = Medium Error/Medium Format Corrupted
Environment
VMware vSAN (All Versions)
Cause
VMDK is using a RAID0 Storage policy and one of the components that make up the object resides on the failed capacity tier disk. As such vSAN can't rebuild the object as there is no redundancy.
Resolution
- Don't use a RAID0 storage policy
- Restore the VM/VMDK from back up
- Replace the failed disk
- See KB VMware vSAN disk encounters medium errors but not failed out by vSAN for more details pertaining to medium errors and setting up alerting to be notified of medium errors.
Additional Information
Feedback
