SQL Transaction Log Corruption since update to vSphere 6.0

[Howarjj]

Ever since we have upgraded to vSphere 6.0, we have been getting corrupted trans log errors while taking transaction log backups of a SQL instance running on a VM. The error message from SQL is: Backup detected log corruption in database Department. Context is FirstSector. LogFile: 2 'F:\Sqldata\Department_log.ldf' I can work around this by setting the SQL DB to Simple Recovery Model and then back to Full and taking a new full backup followed by a trans log backup thereby restarting the chain. This error is happening every night and appears to be kicked off by our nightly VM level backups taken by CommVault which takes a snapshot, backs up the VMDk and then removes the snapshot. However, I can replicate this problem without our backup solution by doing a storage vMotion followed by an in-SQL backup of the transaction logs which will error out with trans log corruption. This was not occurring prior to the upgrade to vSphere 6.0 from 5.5. I have ensured that the VM Tools are up to date as well as the VM Hardware version. Hopefully someone has some insight into what could be going on here as I am stumped and would hate to go backwards to vSphere 5.5.

Relevant VM and vSphere info

VM

     Windows 2008 R2

     SQL Server 2008 R2 - SP3 (10.50.6220.0)

     VM Tools 9536

     VM hardware v11

vSphere

     C7000 Blade Enclosure with 3 x BL460c Gen8

     VMware ESXi, 6.0.0, 2494585 (Custom HP image)

[nickirwin99 · Accepted solution]

Get your vSphere 6.0 Servers patched ASAP.

The same thing happened to us, we had several SQL Server Databases that were becoming corrupted.

It's a known issue with vSphere 6.0 that has since been patched (I think it was related to snapshots and CBT).  The latest vSphere 6.0 patch/build is "ESXi, 6.0.0, 2809209"

[Howarjj]

This seems to have fixed the problem. I was originally hesitant to patch it all up in case we had to roll back. Thank you!

[rajeshkannank]

Hi All,

This post helped me to fix the same issue in our environment. However, VMWare team did not find and helped.

Thanks!!!!

[LDHurst]

We moved our SQL Server 2008 R2 to a new VM 6.0 last week, got this same issue Tues. & Thurs. morning this week.  Tlog backups occur every 30 min from 7am-7pm, full backup at 8pm.  When this error occurs it's on the 7am tlog backup.  What version did you patch it to?  Thanks!