Hi all,

I've a strange issue in my brand new Windows Server 2012R2, let's describe the environment:

Hypervisor = VMware ESX 5.5

Guest OS = Windows 2012 R2 Standard Edition

I've configured the guest OS with several disk for a SQL Server installation, each disk are presented as "pure" virtual disk to Windows that means no RDM (pass through disk).

After a reboot I've discovered that SQL Server has failed to come online, on a brief analisys, all the disks except system drive (C:) and a volume dedicated to swap file (D:) have not come online. They are present in disk management, but they are offline.

I've manually brung online those disk, than reboot and all returned to normality. Subsequent reboot have generated no issues.

In the system log I have no error message except that one of SQL that fails to came online due to the lack of disks...

• Changed type Elton_JiMicrosoft contingent staff, Moderator Wednesday, September 17, 2014 8:03 AM

Hello,

this sounds for me as a ESXi problem and the configuration how the disks are mounted into the VM. Did you already ask in the VMWare support forums?

Hi Meinolf,

first of all thank you for your reply.

I'll investigate with the VMware team to search for any implication of the hypervisor. But I have a doubt: if the virtual machine boots and can't bring his disks online for a reason related to the hardware (i.e. virtua hardware presented by the vmare hypervisor) why there is no evidence in the log?

Maybe I'm seraching in the wrong log? I've serached "system" log and found no entry of warning or error when the operating system has booted, the only error is the MSSQLSERVER service that fails due to the lack of disks.

Hi Andrea,

Did you try to create a new VM on VMWare and attach these disks of SQL virtual machine ?

Best Regards

Elton Ji

Hi Andrea,

have you found any root cause for this? I ran into similar issue today

Hi Andrea,

I did experience this exact same issue last week. Also, VMware ESX 5.5 (hardware VM version 8), also Windows 2012 R2 and SQL (version 2012R2) ; the disks that remained offline were the Database and Log disks. Manually bringing the disks online and restarting SQL services fixed this, but I am reluctant to do a spontaneous reboot and find myself caught in the same situation again.

Did you find anything? What hardware level are/were you using?

Cheers.

Hi Andrea,

I did experience this exact same issue last week. Also, VMware ESX 5.5 (hardware VM version 8), also Windows 2012 R2 and SQL (version 2012R2) ; the disks that remained offline were the Database and Log disks. Manually bringing the disks online and restarting SQL services fixed this, but I am reluctant to do a spontaneous reboot and find myself caught in the same situation again.

Did you find anything? What hardware level are/were you using?

Cheers.

I figured it probably has something to do with the way the disks are re-added to the OS during boot time and de SQL service trying to start before that time, which could be a problem since the mdf's and ldf's are on those disks. But why they remain Offline is still unclear. Anyway, after several reboots, the issue has not surfaced again. If it does, I will try to slow start the SQL service, and give the OS a bit more time to add back the disks during boot time. Hopefully this would help someone with a similar issue.

Hi CS,

kindly change SQL related Service from Automatic to Automatic (delayed start) and reboot the server & try. it may resolve your issue. 

This happened to me too. Except with server 2008 R2 and exchange 2010. My servers replicate offsite using zerto and I was testing failover and noticed the drive D was offline. Brought it online then rebooted and everything was fine.

So I just build a server 2012 R2 Datacenter edition VM on my production servers, gave it a second disk, initialized it, formatted it and assigned it the drive letter D. Rebooted after installing 3 windows updates and it was offline. There is nothing on this server at all. It's a fresh install.

I'm running ESXi 5.5 build 2302651 with hardware version 8 and VMware tools version 9.4.11 build 2400950.

From my experience when you see this kind of stuff happen in multiple situations it will happen again. My worry is what if one of my hosts decides to die and HA kicks in and the disk is offline when it boots.

Anyone have anything to add to this?

Edit: My DR site where I tested my failover is DAS and my production site is iscsi. So I don't think the storage type matters.

Edit 2: The SCSI controller type is Paravirtual for all my VM's.

• Edited by jlongjr Monday, March 02, 2015 6:55 PM

Can you validate the esx log and your store location log. A timeout or a storage problem can prevent the disk to come online.

I seen such symptom when the esx boot and some disk are stored on a SAN that did not mounted enougth fast to get attached to the VM. Thus the error.