Groups | Blog | Home
all groups > sql server clustering > june 2005 >

sql server clustering : Migrate DTC


Hidayat via SQLMonster.com
6/15/2005 12:00:00 AM
tks for your prompt reply.

Another question i would like to ask is do we need different virtual ip for
MSDTC & SQL cluster? Can we use the existing virtual ip that we configure for
MSCS?

Tks

--
Hidayat via SQLMonster.com
6/15/2005 9:05:30 AM
Hi guys,

I have an existing sql 2k enterprise server and we're going to configure it
for server clustering. We have custom made DTC package and most of the
application use it very often.

My concern is how do we migrate this custom made DTC package to the new
environment(sql cluster)?

Appreaciate ur guys help.

Hidayat

--
Rodney R. Fournier [MVP]
6/15/2005 10:01:58 AM
I maybe missing something; due my true lack of understanding MSDTC, but I
don't think you have anything to migrate. MSDTC as far as I know, is just a
log file, and a new install will use a new log file. This should be a walk
in the park.

Cheers,

Rod

MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering Website
http://msmvps.com/clustering - Blog

[quoted text, click to view]

Rodney R. Fournier [MVP]
6/15/2005 11:49:56 AM
If MSDTC has heavy usage then give it its own Group - with an Virtual IP,
Virtual Name, Physical Disk, and the MSDTC resource. If it is not a heavy
hitter, then don't worry about it.

Cheers,

Rod

MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering Website
http://msmvps.com/clustering - Blog

[quoted text, click to view]

Rodney R. Fournier [MVP]
6/15/2005 9:35:35 PM
You can share the space. But if it is heavily used that maybe a bad idea.
Everything about a database is the same, restore, attach, etc. The Books
Online are wonderful and full of great information.

Cheers,

Rod

MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering Website
http://www.msmvps.com/clustering - Blog

[quoted text, click to view]

Hidayat via SQLMonster.com
6/16/2005 12:00:00 AM
Tks for the explaination. Just to confirm with u guys. Let's say that i have
2 node(A & B). I install and configure SQL cluster on Node A. Must i install
the SQL cluster on Node B?

--
Hidayat via SQLMonster.com
6/16/2005 12:00:00 AM
Tks a million guys. Let's say that i have 5GB each on 2 servers. Should i
enabled /3GB or /PAE or AWE? What is u guys best recommendation. I'm going to
configure clustering this Saturday and will be in production.

This clustering is new to me so tks for your patient in answering my question.


U guys have been very helpful.

Tks.

--
Hidayat via SQLMonster.com
6/16/2005 1:07:12 AM
Can the physical disk for MSDTC be the same(sharing) for the SQL cluster
physical disk?

In SQL cluster environment, how is the procedure of restoring the database?
Is it the same as normal SQL restore using it's own utility. As i know the
data will centrally reside in the SAN(EMC) storage.

Tks

--
Geoff N. Hiten
6/16/2005 8:44:02 AM
Clustered SQL servers look and act just like normal SQL servers for amost
every purpose, including backup and restore. There are some extra steps for
things like restoring master, adding disks, and adding/dropping cluster
nodes but these are well documented. From the client's point of view, they
look and act exactly like any other SQL server.

Geoff N. Hiten
Microsoft SQL Server MVP


[quoted text, click to view]

Rodney R. Fournier [MVP]
6/16/2005 10:03:36 AM
The SQL install is cluster aware. Create a cluster on the first node and
have the second one join it. Leave both machine powered on and run the SQL
installer. The SQL Installer will ask you what nodes to install SQL 2000 on,
it will then install on both nodes for you. You need to follow this for
whatever service pack you like (3a or 4).

Cheers,

Rod

MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering Website
http://msmvps.com/clustering - Blog

[quoted text, click to view]

Geoff N. Hiten
6/16/2005 11:04:08 AM
The SQL installer, including Service Packs and Hotfixes, is cluster-aware.
You install to the cluster, starting on the node that is hosting the disk
resource. The SQL installer handles the remaining node(s) appropriately
behind the scenes. Same for SPs and Hotfixes. There are special techniques
for applying binaries to a replacement node, but they are documented in BOL
and the SP readme file.

Geoff N. Hiten
Microsoft SQL Server MVP

[quoted text, click to view]

Rodney R. Fournier [MVP]
6/16/2005 12:15:44 PM
Great question. You need to 1) talk to you hardware vendor and find out if
/PAE or /AWE are even supported. 2) test the /3GB with whatever the vendor
supports, you may not see any difference or worse, things could get slower,
3) I am guessing you will only need /PAE to see the 5 GB of memory, but this
is a guess - testing will tell all.

Lastly, you are on a very aggressive schedule. It sounds like you need more
testing time to me BEFORE you go live. This being your first cluster, can
you restore it or a node in the case of a failure? Do you fully understand
how to handle service packs, etc. I am not trying to beat you up, you have a
learning curve ahead. After you in production is really not the way to
manage and maintain your Highly Available SQL Failover Cluster.

Just my two cents...

Cheers,

Rod

MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering Website
http://msmvps.com/clustering - Blog

[quoted text, click to view]

Tom Moreau
6/16/2005 1:28:30 PM
I can back that up. Clusters are finicky things and can be troublesome to
set up. Make sure you have plenty of time for the job.

--
Tom

----------------------------------------------------
Thomas A. Moreau, BSc, PhD, MCSE, MCDBA
SQL Server MVP
Columnist, SQL Server Professional
Toronto, ON Canada
www.pinpub.com
..
"Rodney R. Fournier [MVP]" <rod@die.spam.die.nw-america.com> wrote in
message news:uKuxIcpcFHA.4040@TK2MSFTNGP10.phx.gbl...
Great question. You need to 1) talk to you hardware vendor and find out if
/PAE or /AWE are even supported. 2) test the /3GB with whatever the vendor
supports, you may not see any difference or worse, things could get slower,
3) I am guessing you will only need /PAE to see the 5 GB of memory, but this
is a guess - testing will tell all.

Lastly, you are on a very aggressive schedule. It sounds like you need more
testing time to me BEFORE you go live. This being your first cluster, can
you restore it or a node in the case of a failure? Do you fully understand
how to handle service packs, etc. I am not trying to beat you up, you have a
learning curve ahead. After you in production is really not the way to
manage and maintain your Highly Available SQL Failover Cluster.

Just my two cents...

Cheers,

Rod

MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering Website
http://msmvps.com/clustering - Blog

[quoted text, click to view]

Geoff N. Hiten
6/16/2005 1:31:08 PM
Clustered SQL Server do's, don'ts, and basic warnings
http://support.microsoft.com/default.aspx?scid=kb;en-us;254321

I also agree with Rodney. You need to get your head wrapped around
clustering befoer going live or you will be lowering your system
availability, not increasing it.

Geoff N. Hiten
Microsoft SQL Server MVP.

[quoted text, click to view]

Mohd Hidayat Had Nadi via SQLMonster.com
6/19/2005 8:06:29 AM
hi guys,

Tks a lot for your time and patience in answering my question. I managed to
get the 2 sql node up and runnning. It is now in production already.

I guess i have the beginners luck. hahaahaa..... Appreciate ur guys advice in
doing a pilot testing 1st but i guess time is an issue.

Tks guys.

--
AddThis Social Bookmark Button