amitrao2009ce
Contributor
Contributor

How to svmotion MSCS clustered VM's using VMDK's configured in physical bus sharing mode (vCenter Server DB nodes)

I have a situation here that I am facing and need assistance guidance from experts in this community.

I have two VM's running Microsoft Windows Server 2012 R2 edition and configured MSCS suing shared VMDK's in physical bus sharing mode. These VM's are running Microsoft SQL Server 2012 and clustered database nodes for my vCenter Server 5.5 (and ESXi hosts 5.5 U3). My VC is not-clustered by my VC database nodes are MS clustered.

My intention is to storage migrate VM disk files to SSD storage from non-SSD storage (current location).

Please suggest suitable way to do so.

Thanks in advance.

Regards,

Amit

0 Kudos
4 Replies
rcporto
Leadership
Leadership

See that another thread: Re: Live Storage Vmotion VMs with shared VMDK

---

Richardson Porto
Senior Infrastructure Specialist
LinkedIn: http://linkedin.com/in/richardsonporto
0 Kudos
amitrao2009ce
Contributor
Contributor

Thanks Rich, I have seen this link before but that's not helping much and do not provide complete clarity. I am looking for clear step way to this problem?

0 Kudos
rcporto
Leadership
Leadership

Basically what you need to do is shutdown the virtual machine participating in SCSI bus-sharing and perform the cold migration.

---

Richardson Porto
Senior Infrastructure Specialist
LinkedIn: http://linkedin.com/in/richardsonporto
0 Kudos
amitrao2009ce
Contributor
Contributor

To be honest, I did that and that all created a hell lot of issues (it was done day before yesterday).

My team did not pay much attention to the point that both VM's were MSCS clustered using shared vmdk's and what they did, shutdown VC DB node 2 and started storage migration (which went successful) and VC DB Node 2 was powered on and MS failover cluster is/was showing all resources online and Green. Now comes the twist in tail, when MS clustered resources are moved from VC DB node 1 to VC DB node 2, the vCenter Server Services go unresponsive and when you look at in Microsoft SQL Management Studio the DB instances on VC DB node 2 shows into suspect mode (Emergency mode) and VC services never come online.

Now when you back migrate resources to VC DB node 1 the vCenter Server services start responding and everything comes back to Normal. That's very strange and now my only intention is to reverse the process and start from there so that's why I'm first attempting to understand svmotion question so that it can be done better and safest way possible.

0 Kudos