In article <bhrmi4$f0m$05$1_at_news.t-online.com>, stepken_at_little-idiot.de
said...
> Hi Billy !
>
> There is a transaction manager, which keeps its eye on all timstamps of
> all inserts/updates and even on timestamps of transactions itself, to
> get all data reconstructed in the right way. In SQL Server you can roll
> back and forward all transactions, right ? Same does MVTO manager. In
> merge replication you have a merge controller, who takes care, that all
> data from all slaves are synchronized correctly into the master, correct
> ? Don't forget that we have versions. Without MVCC and additional
> columns with timestamps, it doesn't work. Same in SQL Server and ACCESS
> (Master-Slave replication). Without hidden columns, it doesn't work.
> MVTO is much more advanced, than merge replication...
So if two transactions are committed to the medical system at the same
time, one marking a patient as dead and the other one saying the patient
is alive, does the database not have to pick a winner? Sure there will be
transactional history on the record (assuming a history logging system
like most medical ones are) but how can we assure that they hit in the
right order?
Received on Mon Aug 18 2003 - 19:18:14 CDT
