From the outside, a cluster failover looks like a stop and restart of the
SQL server, as such any transactions are rolled back/forward and all
connections are closed. Client applications need to be able to handle this
transparently to the end user. This should be part of any well designed and
built client application anyway.
--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
"CD" <mcdye1 (AT) hotmail (DOT) nospam.com> wrote
Quote:
At our center we were under the impression that with a sql cluster system
the app would be up if a node failed and rolled to second node. My
question
is specifically if in this environment should the client die and have to
be
restarted. We had the mind set that the user would not see any issues
expect for a short burp and not having to restart the app.
TIA
CD |