Re: Replication in databases
From: paul c <toledobythesea_at_oohay.ac>
Date: Mon, 01 Sep 2008 23:41:05 GMT
Message-ID: <l4%uk.11265$nu6.8261_at_edtnps83>
>> Hello,
>>
>> I have a fundamental doubt regarding replication in databases.
>>
>> If the isolation level is set to Serializable, why cant we just ship
>> the transaction statements from the master to the backup and replay
>> those transactions. Why do we adopt complicated solutions like log
>> shipping and all such kind of things for replicating databases.
>>
>> Thanks,
>> Srinivas
Date: Mon, 01 Sep 2008 23:41:05 GMT
Message-ID: <l4%uk.11265$nu6.8261_at_edtnps83>
Bob Badour wrote:
> M2Y wrote: >
>> Hello,
>>
>> I have a fundamental doubt regarding replication in databases.
>>
>> If the isolation level is set to Serializable, why cant we just ship
>> the transaction statements from the master to the backup and replay
>> those transactions. Why do we adopt complicated solutions like log
>> shipping and all such kind of things for replicating databases.
>>
>> Thanks,
>> Srinivas
> > The statements might arrive at the slave in a different order and the > slave might serialize the transactions differently--for a start.
from my own experience, another reason is that people who are charged with implementing physical services don't think big (and often are ignorant of product semantics). can't remember how many times I've seen an unstated requirement for secondary mostly read-only db's that can tolerate being minutes or even hours out-of-date. Received on Tue Sep 02 2008 - 01:41:05 CEST