Upgrading an ESXi host fails with the error: Could not stage image profile. Error in purging old directory: [Errno 5] Input/output error:
'/locker/packages/var/db/locker/bulletins'
The following errors are seen in /var/log/esxupdate.log
YYYY-MM-DDT23:32:08Z esxupdate: 2112967: vmware.runcommand: INFO: runcommand called with: args = '['/usr/lib/vmware/vob/bin/addvob', 'vob.user.esximage.install.stage.error', '(Updated) ESXi-7.0U3g-20328353-standard', '(\'/locker/packages/var/db/locker\', "Error in purging old directory: [Errno 5] Input/output error: \'/locker/packages/var/db/locker/bulletins\'")']', outfile = 'None', returnoutput = 'True', timeout = '0.0'.
YYYY-MM-DDT23:32:09Z esxupdate: 2112967: esxupdate: ERROR: An esxupdate error exception was caught:
YYYY-MM-DDT23:32:09Z esxupdate: 2112967: esxupdate: ERROR: Traceback (most recent call last):
YYYY-MM-DDT23:32:09Z esxupdate: 2112967: esxupdate: ERROR: File "/tmp/esx-update-2112967/lib64/python3.8/site-packages/vmware/esximage/Database.py", line 407, in _Commit
YYYY-MM-DDT23:32:09Z esxupdate: 2112967: esxupdate: ERROR: File "/lib64/python3.8/shutil.py", line 743, in rmtree
YYYY-MM-DDT23:32:09Z esxupdate: 2112967: esxupdate: ERROR: File "/lib64/python3.8/shutil.py", line 599, in _rmtree_unsafe
YYYY-MM-DDT23:32:09Z esxupdate: 2112967: esxupdate: ERROR: File "/lib64/python3.8/shutil.py", line 597, in _rmtree_unsafe
YYYY-MM-DDT23:32:09Z esxupdate: 2112967: esxupdate: ERROR: OSError: [Errno 5] Input/output error: '/locker/packages/var/db/locker/bulletins'
VMkernel reports a SCSI sense code of 0x3.
Log file: /var/log/vmkernel.log
YYYY-MM-DDT22:16:27.989Z cpu2:2098115)ScsiDeviceIO: 4167: Cmd(0x45b95771bd88) 0x28, CmdSN 0x55abf from world 2123254 to dev "naa.######" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x3 0x11 0x0 Medium Error, LBA: 2810240
YYYY-MM-DDT22:16:30.980Z cpu2:2098115)ScsiDeviceIO: 4167: Cmd(0x45b957714088) 0x28, CmdSN 0x55bc0 from world 2123254 to dev "naa.######" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x3 0x11 0x0 Medium Error, LBA: 2810240
vSphere ESXi
Input/Output errors usually occur due to underlying storage issues.
The /locker folder resides on the OSData partition. A failing boot volume will cause the I/O error.
Medium Error indicates that a SCSI read (Command 0x28), failed on the device due to hardware error.
Investigate the cause of the I/O error from the VMkernel logs.
The below error indicates that a SCSI read (Command 0x28), failed on the device due to hardware error (Medium Error)
Cmd(0x45b95771bd88) 0x28, CmdSN 0x55abf from world 2123254 to dev "naa.######" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x3 0x11 0x0 Medium Error, LBA: 2810240
If Medium errors are reported,
For more information, see Interpreting SCSI sense codes in VMware ESXi.