I recently faced an issue with backing up and restoring a Linux VM. The issue happens with a Ubuntu 16.04 VM that has a the Veeam Linux agent installed (deployed via Veeam B&R). There is one backup job backing up this particular VM and a backup copyjob to transfer it offsite.
While doing either an Instant Recovery or VMDK export of the backups, the exported images are corrupted and can't boot successfully due to several shared libraries having ELF Header mismatches. libsystemd is one of those shared libraries which end up corrupted/changed. Even apt can't be used as it also relies on this particular library. While fixing this issue is possible, its not an option for desaster recovery.
The issue does not happen on the production side, we didn't detect any disk corruptions or ELF header mismatches there. The corrupted state occurs after the backup only. Several new full backups have been tested, all of them showed the same symptoms.
We checked several logfiles for errors already, among them the Linux agent's ones. No obvious error messages are getting displayed here.
This is a customers VM running on an ESXi host. How to proceed, are there more logs to check? Is this a common issue? Thanks in advance!
Cheers!
(case id is #05315411)