Yep, I also think so. I'm currently developing a small prototype for this kind of transparent proxy, which I'll post here when it's stable...
Tanel. > Tanel, > > I think this is a good solution, provided the application can handle > two phased commit protocol across both the databases, else there > could be orphan records on one or both these databases. > > Murali Tanel Poder <[EMAIL PROTECTED]> wrote: > When you want true 24x7 without compromises, then you have to step > closer tothe client anyway.This means, you have two databases for > example and your app servermultiplexes all transactions to both > ones.This should be faster than sync standby or sync replication, > because appserver can send requests to both databases in parallel, > unlike withstandby/replication where database itself resends the > requests to otherdatabases, making the "chain" of changes > longer.Tanel.----- Original Message ----- To: "Multiple recipients of > list ORACLE-L" Sent: Wednesday, December 10, 2003 11:24 PM>>>>> Hi,>> > Unfortunately I'm gonig to add the negative view, like several > others> have...>> True 24x7x365 (good pick-up Pete on the 7 year > thing) will! > be > limited by> much more than database and operating system > availability. We just did a> major software upgrade last weekend and > part of the upgrade involved the> conversion of 250+ million records > in the database - that takes time no> matter what. Yes, with > unlimited budget and time constraints we could get> the outage down > to nothing but at the end of the day it's easier for the> business to > manage an outage.>> Our system was offline for a total of about 10 > hours yet traffic drives on> our tollroad all the time so: The > roadside is designed to backlog> transactions for several days, our > system has capacity to catch upbacklogs> fairly fast (within a day we > had caught up again), we have an alternative> front-end system that > can backlog feeds, and finally we designed our> conversion process to > do as much as possible before the outage.>> We are also on 8.1.7 > enterprise and! > don't > use OPS/RAC - instead we havethe> alternative processes in place to > ensure the business can function.> Perhaps your company can consider > a similar alternative? As others have> said - it can be VERY > difficult to remove every possible outage and it can> be much easier > to manage a small outage every few months.>> Regards,> Mark.>>>>> > "Tracy Rahmlow"> !!!!!
