September 23, 2004 at 5:44 am
Hi!
I need to make an active/active cluster out of two servers. One of them is critical and has huge number of published tables, so I want it to be available the sooner the better. What is the best way to do this?
When moving critical servers I practiced this aproach:
1. Stop sql-server service
2. Copy all databases file on a new server
3. Shut down the old one
4. Name the new the same name and install MS SQL Server 2000 on it
5. Stop the sql-sever service and replace the its databases files with the ones copied from the old server
6. Start the sql-server service
Since one of the sql server nodes will have to be renamed the way like: node2\second, the above aproach will not work for the second server. But will it work for the first? Are there any subtleties concerned with possibly different way a clustered MS SQL Server works?
Thanks.
September 24, 2004 at 9:37 am
Do you not have shared storage between the two servers?
September 24, 2004 at 11:00 am
I'm not sure if i understood your question correctly but if you are trying to make active/active out of two existing servers having sql installed, you can follow the steps below
1) You need a shared disk. configure your server with a shared disk.
2) you will also need a heart-beat cable. A cross-over cable. You will also need two I.P. and network nameS for virtual sql I.P. and virtual network name and cluster I.P. and cluster name.
3) Install the cluster software.
Once the above is verified,
4) Install SQL on the cluster. You can have SQL Running as a stand alone instance.
5) Once SQL is installed, you can detach all user databases from the stand alone installation, copy the files to the installation folder of the clustered server and then attach the files. Do not copy the system databases from the stand alone server to the clustered server.
6) you can copy the logins over by using the script below.
http://support.microsoft.com/view/tn.asp?kb=246133
7) You can now un-install SQL Server on the stand alone server or disable it. Be careful not to touch the virtual instance.
September 26, 2004 at 4:01 pm
I have new hardware for the cluster, so I do not need to install several instances on the same server. The hardware has shared storage. All the hardware questions like shared storage, heartbeat etc. are clear to me.
5) Once SQL is installed, you can detach all user databases from the stand alone installation, copy the files to the installation folder of the clustered server and then attach the files. Do not copy the system databases from the stand alone server to the clustered server.
The problem I have is the huge number of replications on the servers. Therefore I can not detach and attach databases. (Recreating all the replications, jobs etc. all other again is not an option: there are a lot of triggers on subscribers and the number of replicated tables is huge. I will have to stop the whole system for this.) As for detaching I can work it around by setting category column in sysdatabases table, but as for attaching I am not sure this will work same way. I need to copy at least some system databases (perhaps msdb and distribution) I can probably fake the master database by creating similar structured databases, stopping the server and replacing the databases files. So I need some kind of copy and past aproach like the one I used when moving critical servers to new hardware.
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply