Error Message

0xe000fe36 - Corrupt data encountered

This is caused by a change in the VSS stack inside the VM after the AWS backup job has taken a snapshot and started the backup operation. The change in the VSS stack is actually caused by a snapshot request against the VM itself being scheduled to take place during the AWS backup job. This snapshot request causes a call through VMware Tools (inside the VM) to run pre-freeze-script.bat which registers and starts the BeVssProvider service and it is this that modifies the VSS stack and causes the problem against the active VSS snapshot being used by the AWS backup job. 

Notes:

Whilst this can be caused by scheduling a Virtual Agent/VMware Agent backup job within Backup Exec to start during the Agent for Windows backup, it has also been seen that 3rd party products managing the VMware environment may also be scheduled to take snapshots of VMs and can therefore trigger the BeVssProvider service start even though the product in question is not Backup Exec. As an example Technical Support have seen support cases where Dell software (supplied with Equalogic Disk Arrays) has been configured to take snapshots at a VM level within a virtual environment that was being backed up by AWS backups. 

It is possible the same situation might effect traditional agent backups of database technologies such as Microsoft Exchange or Sharepoint and the error messages for these environments may not exactly match those mentioned in this article.

Confirm by comparing the timings of the AWS backup job with the content of the Application Event log inside the remote system that no BEeVssProvider operations take place during the timeframe of the backup. Specifically look in the Event log for Event ID 0 from Source BeVssProvider that shows the text content indicating either "Service started/resumed" or "Service stopped"

If a match cannot be identified for such entries in the event logs and the timings in the AWS Job Log then there is a different problem being experienced. 
If a match of such events to the same timings as the AWS job is found then identify which products might be making VMware snapshot requests and adjust the schedules of such requests (or disable them) so that they cannot take place at theism time as the AWS jobs.

In some environments, it may be necessary to run traditional Agent for Windows Servers (AWS) backups of the file systems of virtual machines (VMs) hosted inside of VMware ESXi. Whilst this type of backup should run correctly a situation exists where the AWS job might fail with 0xe000fe36 - Corrupt data encountered.