VCS Service group configuration

[Chukwuemeka]

Hello Guys,

Need some explanation and clarification here.

I have two systems installed with windows server 2008 R2.

The both server currently have two hard drives the C drive and an E drive.

The E drive on the both systems are currently connected via iscsi from a NAS.

Names of the both system are winsvr1 and winsvr2

Both systems are guest virtual machines configured on hyper V.

Name of the E drive on both systems are STU (E:)

Now i have SQL server 2008 installed on both systems.

I also have three virtual network cards with various ip addresses on them on both systems.

The binaries are installed on the C drive while the following were installed on the E drive(MSSQL, data, logs, jobs etc)

Now i have also installed VCS 6.0.1 on one of the systems(winsvr1) and i am trying to configure the service group to start on that system before

replicating the same thing to winsvr2 and then configuring the failover.

See attached also what is started up on my services.msc..... what should be put on manual or stopped.

Now on trying to configure the vcs i am trying to use the tools>configuration wizard and also trying to choose a template and modify the template.

Now i see so much on sql 2008 on the templates which one do i take and what next should i really do to get this service group online and working.

I am basically stuck on what next to do to get the service group started.

[Gaurav_S]

Hello,

Please give us a screenshot of VCS console as to what is running & what is not running at this stage so that we can confirm on what resources need to start exactly or any of them is failing.

I would assume that cluster agent for SQL would be able to online the SQL services.

Have you referred the admin guide for VCS Windows for SQL 2008 ? One of the comment on page 25 says:

Note: If you start the SQL server services from outside VCS, then the SQL resource
will go in an UNKNOWN state, because the VCS agent monitors the computer
context of the services. If the SQL service is not started in the virtual server context
the resource goes in an UNKNOWN state. You must ensure that you start all the
SQL related services from within VCS.

G

[Marianne]

Have you installed Storage Foundation/HA or just Cluster Server?

It is best to perform installation of both cluster nodes at the same time and then use the wizard to perform basic cluster config...

You may want to download the Installation Guides from here:

https://sort.symantec.com/documents/doc_details/sfha/6.0.1/Windows/ProductGuides/ 

Under Installation and Upgrade you will find:

Veritas Cluster Server Installation and Upgrade Guide - English

and 

Veritas Storage Foundation and High Availability Solutions Installation and Upgrade Guide - English

These docs describe requirements as far as Private Network (heartbeats) are concerned and other pre-reqs.

Once basic cluster config is done, use this Guide to install SQL into the cluster and config the Service Group:

Veritas Storage Foundation and High Availability Solutions HA and Disaster Recovery Solutions Guide for Microsoft SQL Server 2008 and 2008 R2 - English

[mikebounds]

You have to install SQL as per the guide http://sort-cdn.symantec.com/public/documents/sfha/6.0.1/windows/productguides/pdf/SFW_HA_DR_SQL2008...

where you install "SQL Server stand-alone installation" on each node and as part of the install you set the services:

On the Server Configuration panel, change the SQL Server services Startup

Type as follows (or use the SQL Server Configuration Manager to change it

after the installation):

■ Set the SQL Server Browser service to Automatic start.

■ Set all other SQL Server services to Manual start.

You should follow this guide for the installs where you install the first node with the database on the shared (E:) drive and for the second node you install database to local drive (C:)

Mike

[Wally_Heim]

Hi Chukwuemeka,

There are a few ways to get the SQL service group configured.  However, using the Templates would not be my first choice.

Try one of these options:

1. Use the "SQL Server 2008 Configuration Wizard" that is located in Start > All Programs > Symantec > Veritas Cluster Server > Configuration Tools.

or

2. Use the Symantec Solutions Configuration Center (this starts automatically after login once the product is installed.)  Go to the Solutions for Microsoft SQL Server and select the High Availability (HA) Configurations (New Server) option.  This will give you step by step what needs to be configured.

As a note, you will need to set your SQL services to manual.  Specifically, you need to set the SQL Server (MSSQLSERVER) and SQL Server Agent (MSSQLSERVER) services to manual start up on both systems.

Keep in mind that if you have problems you can always open a case with Symantec Technical Support and we will be happy to help you.

Thank you,

Wally Heim

[Chukwuemeka]

Thank you guys, i am going through the docs again for more clarification,  but would ask more questions as i go on.

Now i forgot to put it in my questions,

On the second server winsvr2 i have all the installation done on the C drive, would i need to actually uninstall again and do it the same way like winsvr2.

@ Marianne, i installed the entire suite of SFWHA and included is also vcs.

And would get back to the forum soon.

[Wally_Heim]

Hi Chukwuemeka,

You may not need to uninstall on winsvr2.  I'm assuming that you are referring to SQL not being installed the same on winsvr1 and winsvr2.  As long as the basic installation path is the same for the SQL binaries but the Master Database is in different locations, VCS can correct this.  However, if the path to the SQL binaries are not the same on winsvr1 and winsrv2, then you will need to uninstall/reinstall SQL so that the binary path is identical.

If you are talking about SFW-HA is not installed to the same path on all nodes, then I would recommend correcting this because you might get odd behavior as you try to do some cluster operations.

Thank you,

Wally

[Chukwuemeka]

@ Wally.

My current suitation is

On winsvr1, i have everything( binaries, data, logs etc) all installed on the C drive then on winsvr2 i have only the binaries installed on the same path as winsvr1 on C but the (data, logs, jobs etc) are all installed on the E drive.

The winsvr1 is supposed to be the primary site while the winsvr2 the secondary...

Now from the pdf it says

1.  Installing SQL Server on the first system

    Install the SQL Server instance to the local system disk.

    If you are configuring SQL in a non-shared storage environment, install the

    SQL Server database files and analysis service on to the local attached

    non-shared storage (non system drive).

2. Installing SQL Server on additional systems.(i am assuming this would be winsvr2)                                Install the SQL Server instances, database files, and analysis service files on the local system disk.

Note: In a non-shared storage configuration, installation of SQL Server on

additional systems is not applicable

Now i dont understand that because i am using separete hard disk for both servers as their E drives and not a shared hard drive as E.

So what is supposed to be the recommended installation setup for SQL failover over two systems.

I was thinking if you have the binaries on C drive and data, logs etc on E drive on a particular system(winsvr1) the same should apply for the second system(winsvr2)

[Wally_Heim]

Hi Chukwuemeka,

If you were using a shared drive for the E:\ drive, then the SQL wizard when run non winsvr2 would be able to correct the database path on winsvr1.  However, you are not using shared storage, the wizard will not be able to complete and make these changes for you.

If you are not able to run the SQL 2008 Configuration Wizard, then I would recommend installing both winsvr1 and winsvr2 identically.  But I need more details on what you are planning to say that is 100% needed or not.

Were you using separate E:\ drives just for installation or do you plan to use replication or something else to get the data from server to server?  If you are not going to use shared storage in your final configuration, how are you planning to get the data from winsvr1 to winsvr2?

-Wally

[Marianne]

The Manual for SQL config covers the following in Chapter 7:

Chapter 7 Installing SQL Server

About installing and configuring SQL Server ...............................................175

About installing multiple SQL Server instances ...........................................176

Verifying that SQL Server databases and logs are moved to shared storage 177

About installing SQL Server for high availability configuration ...............177

About installing SQL Server on the first system ..........................................179

About installing SQL Server on the second system .....................................181

Creating a SQL Server user-defined database ...............................................182

Completing configuration steps in SQL Server .............................................183

Moving the tempdb database if using VVR for disaster recovery ......183

Assigning ports for multiple SQL Server instances .............................184

Enabling IPv6 support for the SQL Server Analysis Service ..............184

Steps are in short:
Create a diskgroup on Node1.
Create volume(s) for databases and logs and specify drive letter (or mount point(s)).
Install SQL with binaries in Program Files and databases and logs on the newly created volume(s).
Set SQL services to manual, then stop services.
Deport diskgroup and import on node 2. Map volume(s) to same path/drive letter(s).
Wipe or rename folders created by installation on node1.
Install in the same manner as on Node1.
Please bear in mind that you need a separate, small volume for Registry Replication.
Have all Service Group properties ready in a worksheet before you start the service group wizard.

These steps are just a short overview - please go through the documentation.

Proper preparation is the key to successful installation and configuration.

[Chukwuemeka]

Hello wally, Moving the data from winsvr1 to winsvr2 would be done by vvr. Let me just start the installation again on the winsvr2 system try and startup the service group then proceed to winsvr1. Because starting from winsvr1 might be a bit complex because I have all the installation folders, data,logs etc all installed on a single directory which is the C drive and from the pdf this should not be the case for vcs to work.

[Wally_Heim]

Hi Chukwuemeka,

What you are configuring is a Replicated Data Cluster.  It is starting to become a more common configuration.

winsvr1 would need to be reinstalled so that the Master database files are in the same location as used on winsvr2.  I don't really think you need to do anything with winsvr2 from the sounds of the installation there.

Please open a support case if you need more direct assistance with this configuration.

Thank you,

Wally Heim

[Chukwuemeka]

Hello guys this is where i am now and need some clarification.

Winsvr1 and winsvr2 are 2 virtual machines and they are both single node cluster.

No form of storage is being shared here, the replication is going to be done by VVR.

winsvr1 has already SFWHA installed and SQL installed as directed by the admin guide.

Winsvr1 SQL binaries installed on C then other data files installed on E drive as directed on the guide.

As stated earlier they both have other drives connected to them via scsi on the vmware platform.

Now on winsvr2, i have also installed SFWHA and like the admin guide directed i installed both the SQL binaries, data files, logs etc on the local C drive so as not to have the same path with the winsvr1 server.

Somewhere on the admin guide i saw this and need some more clarification.

Note: In a non-shared storage configuration, installation of SQL Server on

additional systems is not applicable. ...

page 181.......SFW_HA_DR_SQL2008_Solutions_6.0.1.pdfSFW_HA_DR_SQL2008_Solutions_6.0.1.pdf

Because in my own scenario i see my storage as non-shared storage configuration.

[Chukwuemeka]

Hello guys,

Thank you for all your comments.

I have been able to sort this out with support team.

Closing this for now.

[Marianne]

Care to tell us HOW it was resolved?

[Chukwuemeka]

Hello Marianne.

Would comment on the resolution soon.

[Chukwuemeka]

Hello Guys,

Basically how it was resolved, 

I started with the template SQLServer2008-VMNSRVGGroup did all the editing making reference to the vcs simulator SQL_DR configuration.

Then with the help of support the missing values were added to the RVG resource and replication of the database was also configured before i was able to start up the sql service group and fail it over to the second node of the global cluster configuration.

Anyone that needs any deeper information can let me know i would be glad to share it.

thanks