The database will work with the available logs and replication records to commit
every transaction that can be committed, rollback every transaction that should
be rolled back, and, worst case, mark as "in-doubt" any transaction which will
need a DBA's help to be figured out (nothing will be marked in-doubt in almost
every circumstance). As in any crash situation, data integrity comes first.
Erik Hendrix wrote:
[color=blue]
> Hey Blair,
>
> What would happen with any open transactions? Since the app. does not need
> to reconnect, I assume that app never knew that the database crashed then
> and moved to another server.
> But any open UOW will be rolled back by the database. So what happens there?
>
> Thanks.
>
> "Blair Adamache" <badamache@2muchspam.yahoo.com> wrote in message
> news:ca9nt3$h62$1@hanover.torolab.ibm.com...[color=green]
> > You're probably best to download the beta (or at least the beta
> > documentation):
http://www-306.ibm.com/software/data/db2/stinger/
> >
> > Before Client Reroute, when a system crashed, its IP address would be
> > assumed by the standy machine (or mutual takeover machine), the app
> > would lose the connection and have to reconnect.
> >
> > With Client Reroute, the connection is transferred with the IP address,
> > so if all goes according to plan, the application need not reconnect.
> >
> > hrishy wrote:
> >[color=darkred]
> > > Hi Blair
> > >
> > > Can you say more about client reroute on stinger..
> > >
> > > regards
> > > Hrishy
> > >
> > >
> > > Blair Adamache <badamache@2muchspam.yahoo.com> wrote in message[/color][/color]
> news:<ca57i1$360$1@hanover.torolab.ibm.com>...[color=green][color=darkred]
> > >
> > >>If you measure the number of machines in the cluster as a measure of
> > >>scalability (i.e. how large a database can be supported, how complex the
> > >>queries can be), my limited and biased experience shows DB2 to be ahead
> > >>of Oracle on UNIX, Linux and Windows.
> > >>
> > >>When DB2 Stinger ships (later this year), which has a feature called
> > >>Client Reroute, Oracle and DB2 will be more or less equal for
> > >>transparent application failover (Oracle's shared disk approach has
> > >>strengths and weaknesses compared to the shared nothing approach which
> > >>has been adopted by DB2 UNIX/Linux/Windows, Informix XPS, Teradata and
> > >>SQL Server).
> > >>
> > >>BTW, in this newsgroup, your question is about to cause a free for all
> > >>similar to throwing a raw steak into a Lions' den.
> > >>
> > >>Mike wrote:
> > >>
> > >>
> > >>>We're starting a project at work moving VSAM to RDBMS. The choice
> > >>>is between DB2 and Oracle. It seems like the Oracle RAC is a better
> > >>>cluster choice with it's share everything rather than the DB2
> > >>>share nothing. Please post some opinions on this and/or other
> > >>>points of difference/intereste between the two DBMS.
> > >>>
> > >>>Mike[/color]
> >[/color][/color]
