Clustering and Transactions

Hi,
When the primary server fails and the secondary starts to take over the job,
what will happen to the transactions opened on the primary node prior to
failure?
Thanks,
Leila
They all get rolled back and every connection is severed. Your client code
should re-try.
Tom
Thomas A. Moreau, BSc, PhD, MCSE, MCDBA
SQL Server MVP
Columnist, SQL Server Professional
Toronto, ON Canada
www.pinnaclepublishing.com
"Leila" <leilas@.hotpop.com> wrote in message
news:OXWJoSdwEHA.3336@.TK2MSFTNGP11.phx.gbl...
Hi,
When the primary server fails and the secondary starts to take over the job,
what will happen to the transactions opened on the primary node prior to
failure?
Thanks,
Leila
|||"Leila" <leilas@.hotpop.com> wrote in message
news:OXWJoSdwEHA.3336@.TK2MSFTNGP11.phx.gbl...
> Hi,
> When the primary server fails and the secondary starts to take over the
job,
> what will happen to the transactions opened on the primary node prior to
> failure?
You need to treat it just like one server was shut down and the other
started up.
Transactions are rolled back (or in a few cases forward) as appropriate.

> Thanks,
> Leila
>

Clustered DB Server

I had prior experience in clustering DB servers. But at that time, the OS did seem to be mature enough to handle clustering. Does any one have more current experience about how clustering sql database servers runs. Faster or slower? Any performance gain at all? Any opinions will be appreciated.Sorry! I made the wrong statement. I meant to say when I did the clustering, the OS did NOT seem to be able to handle it very well. I am looking for opinions from any one who has experience in clustering in a Windows 2000 or 2003(perferrable) environment.|||Windows 2000 seems to have made decent progress in this area, but we have had extraordinary pain with our hardware not liking life. Be forewarned that although MS says it is up and running 24x7, there are some hardware fixes that will bring you down entirely. Notably anything to do with your disk array/cabinet. Also make sure that you place the quorum on its own physical (that's right, physical) device. I know the quorum takes up all of maybe 100MB, but MS suggests this, in order to keep the server running. Basically, if you generate enough I/O on the disk, you can lose sight of the quorum disk, which means the cluster will fall down and go boom.