Go Premium for a chance to win a PS4. Enter to Win

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 217
  • Last Modified:

Designing Failover Clustering for Windows 2012 and recommendation for shared storage.

I'm planning to migrate our servers to Failover Clustering. I'm still in the stage of studying and researching, but need help to go in detail of this process.

Environment:
+2 Physical HP Proliant servers, 2X 6 cores, 32GB ram.
+8 Windows server 2012 VMs in total.
+2 SQL 2012 STD  (Primary and Backup for daily report generation purpose, Over night Primary SQL backupand restore full database backup to Backup SQL), about 50GB database
+1 Exchange 2010 STD, about 80 mailboxes
+2 File servers, about 100GB data

1. The total disk size needed for this environment is about 200GB giving more space, I say 300GB. I have to evaluate what shared storage is suitable for us. Can you recommend any product to accommodate this environment?
*I have no experience with SAN or ISCSI, but have knowledge by learning, also have tested iSCSI in my Windows Failover clustering using VHD.

2. Between, Servers and the shared storage, do you recommend gigabit or 10gigabit switch? Currently we have Dell Powerconnect 5xxx gigabit switches

3. For SQL and Exchange clustering, Do I set up regular VM clustering or others?

4. What about DC? I should not cluster Domain controllers because they replicate each other already. Is it better to have another physical server or configure one of the two HP Proliant servers as a DC?

5. For SQL, I'm not sure if it's good to run as a VM. SQL is currently installed on a physical host, one of HP Proliant servers.

Thank you in advance.
0
crcsupport
Asked:
crcsupport
  • 2
1 Solution
 
Cliff GaliherCommented:
1) For that environment, with no existing shared storage in place, I'd grab a SAS shelf.  Since you already have HP servers, getting HP SAS HBAs and a SAS enclosure that is certified for failover clustering should be easy. The MD3000 series comes to mind if I recall correctly.  A SAS enclosure will be significantly less expensive than a SAN (iSCSI or FC.)

2) If you have storage traffic anywhere on your network (iSCSI or SMB3, for example) it should be *at least* 10GB, if not higher. At 1GB, your network would be a *huge* bottleneck.

3) Exchange does not support being run on a cluster. It has its own HA solution (DAG) which happens to be built on Windows failover clustering, but you can't have the VM itself failover or problems arise. Same goes with SQL. Use their solutions, not clustered VMs for those workloads. This is covered in much more detail in the system requirements for those products.

4) Clustering a DC is *not* supported. And running anything besides the hyper-v role on physical hardware is also very bad. You'll either want a physical DC or put one DC on each node (so a DC is always available) and configure them to never fail over.

5) That depends on the workload and your performance requirements. SQL is certainly supported in a VM. Whether that's good for you....we don't have enough information to answer that at all.
0
 
crcsupportAuthor Commented:
If I put more SAS drives to enclosure of one of the two HP servers, how do I implement CSV for the clustering?
Physically the storage will be at one of the servers will make no failover?
Do you mean an external SAS enclosure and connect with the two servers?
0
 
crcsupportAuthor Commented:
I see, directly attached storage, MD3000. Thanks.
0

Featured Post

Problems using Powershell and Active Directory?

Managing Active Directory does not always have to be complicated.  If you are spending more time trying instead of doing, then it's time to look at something else. For nearly 20 years, AD admins around the world have used one tool for day-to-day AD management: Hyena. Discover why

  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now