My Colleague is on hospital right now and he´s the best one handling the NetApp here. He won’t be back for 8 weeks so I have to get a new 2-node Active/Passive SQL Cluster up and running on my own..
What I have is Windows 2008 Server ENT, SQL Server 2008 R2 STD. And VMware ESX 4 with vSphere Virtual Center. And ofc the NetApp Filers. I have today a big VMware environment up and running with some 400 VM´s on the NetApp, but no Windows Clusters.
I would like to have some guidance where it comes to installation order of all components and some on configuration of those:
Microsoft Cluster Services
iSCSI initiator ( well it´s on the control panel )
NetApp Host Configuration Utility
NetApp SnapManager for SQL
Microsoft SQL Server
And where and from what I create and bind the LUN:s, resources in the right order for the cluster services.
I haven’t found any real help on the internet for this scenario, and since my colleague is on hospital, everything has to happen right now, a thousand things to do..
The only thing I found that I have been thinking about is this scenario: https://communities.netapp.com/community/netapp-blogs/msenviro/blog/2011/02/18/install-order-for-hyper-v-clusters But i Don’t know if I need to have MPIO, Host Configuration Utility and so on.
Can you help me with some guidance please
Hi and welcome to the Community!
- are you talking about clustered SQL nodes running on physical hardware, or as VMs on top of vSphere?
- what version of SMSQL are we talking about?
- what version of SnapDrive?
So this is a very tricky subject then.
For a starter, 'proper' iSCSI (via VMware stack) is not supported for Microsoft clusters:
VMware is hinting a workaround in vSphere 5 by using software iSCSI initiator inside a guest:
But this is not the recommended setup on the NetApp side - all recent best practices suggest using VMware iSCSI stack, which brings us to square 1: MSCS not supported.
A a bit risky scenario - especially if you have no extensive experience with NetApp (if I understand your initial post correctly).
You can easily find some best practices documents here (limit search to Technical Reports):
but I don't think any of them will match your exact requirement.
For this cluster, should I create separate volumes for Quorum, MSDTC and Data ? Like this ?:
Volume1, will be for LUN Quorum
Volume2, will be for LUN MSDTC
Volume3, will be for SQL Databases and logfiles
And question 2, shuld I Snapshot Quorum and MSDTC ? I guess not, and this is the case of not placing them on the same volume as the Snapped Volume3 for SQL Databases.
When configuring SnapManager for SQL, should I install and configure this on both cluster nodes ? Anything special to consider ? Should each cluster node have it´s own SnapInfo-LUN, or should I have one and put it in the SQL Resource Group ?
You must install SnapManager SQL on all nodes. Each instance should have as a minimum
DB Log Drive
System DB & Log Drive
So yes depending on your setup you might need to have more SnapInfo Disks. SM SQL can also deal with volume mountpoints if you run out of drive letters.
Lun1: User Database and log
Lun2: System Database and log
Lun3: Temp Database and log
When configuring SMSQL, should I enter the Cluster name for SQL or each node in the initial screen
You need to split Lun1
Lun1 User Database
Lun2 User Database Log
Lun3 Syste, Database and Log
Lun4 Temp Database and Log
Lun5 NetApp SMSQL SnapInfo
When configuring backup jobs enter the cluster name.