Hi !

I've finished setting up my Virtual Cluster Server on
Windows 2003 server. My cluster System contains two
nodes on Windows 2003, the storage share bus's hardisk
is partitioned as Q disk (1Gb-used for Quorum) and P Disk
(202GB - used for Cluster Data) with Raid ADG Fault
tolerance. And My group resources contain "Cluster group
have cluster IP Address-Network name, and Physical Disk
(Q) resource," and Group 0 (with Disk P Resource into).I
tested . the valid verfied failover of cluster system.It's
ok.

I've finish setup my SQL Virtual Cluster on two node
Because I think the config order to create MSDTC resoucre
of SQL 2000 Virtual Cluster on my cluster system doesn't
correct. So I want to settup SQL 2000 Virtual Server
with MSDTC again. Must I have to setup it again? But
before setting up, I'd like to ask
you about the correct installation order.

After finish my cluster service sinstallation, I begin
start setting up SQL 2000 Virtual Cluster .

The first node is owner of all resource cluster.
First, I begin to install the Single node(durring setup,
possible owner is added to first node) to add the SQL
Virtual Cluster Server on the first node (current, The
first node is owner all the resource's cluster and the
second node is shutdown).
I setup by the SQL 2000 Enterprise CD.The first
installation is complete successfully
The Group 0 now has new more five SQL Resources. and all
of them is online status.

Second I turn on the second node (the first node is still
turning on OS and it is also owner of all resource cluster)
On the second node, I add the second node to SQL virtual
cluster as this same above installation (durring setting
up, possible owner is added to the second node). And the
second installation is also complete successfully.

I also test the failover of my SQL cluster on my cluster
system , it's ok. And
the SQL server run on two server is also ok.

But my database need MSDTC to run SQL (it used
DistributeTransaction Cordinates service).
So, I create a MSDTC Resource (type Distribution
Transaction Data) in Group 0 as "Microsoft Knowledge Base
Article - 30160 guide". And I think if i create MSDTC
resource on this group 0, it doesn't affect the peformance
issue, because the size of disk P is enough big (202Gb).
but the result it is failed status (not online) when I
bring MSDTC Resource is online. And
the Distribute Transaction Cordinates service does not
still Start.
(As I know, if setup successfully, This service must start
only on the curren owner node, and this service also start
on the other node when the current owner node is fail over)


So I wonder that my Installation order with MSDTC of SQL
2000 virtual cluster is correct? Must I create MSDTC
Resource on the first node right after the single node
SQL cluster installion on first node finished. Is it
right ? And after I
finish setting up the SQL virtual Server on the second
node, What things Can I do to
bring MSDTC online and fail over right on my cluster SQL
system.
If can, please help me the correct order installation SQL
2000 Server with MSDTC
on Windows 2003, (if can, please help me to do as your
detail steps guide)

Thanks you very much.I've look forward to your replies.
Regards.
Huy Nguyen

Please email to me at address "nhm_huy (AT) yahoo (DOT) com" and
huy-hm_nguyen (AT) hptvietnam (DOT) com.vn

Wow. You must like doing things the hard way. Here is how it is supposed
to work.
This is assuming your cluster is installed and running.

Bring all nodes online and make sure the cluster resource group can move to
all nodes.
Run comclust.exe on all cluster nodes. This will create a MSDTC resource in
the cluster group. You only need one MSDTC resource for your entire cluster
but you must run comclust on each node.
From the node that currently owns the disk resource you want to be the
default data location, install SQL Server Enterprise Edition. SQL will
install to up to four nodes with one install. The install will create a
virtual server for your SQL Instance. Test moving the virtual server to
each allowed node.
Move any additional disk resources into the SQL Instance resource group and
set dependancies as needed. You can move log files for existing system
databases as per BOL.
Run any additional SQL installations to create more virtual servers as
necessary
Apply service packs and hotfixes to all Virtual instances.
Clean up resource group names and allowed/preferred node lists.
Backup OS and system databases for each instance.

You are now ready for user data.

--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
CareerBuilder.com


"Huy Nguyen" <nhm_huy (AT) yaoo (DOT) com> wrote