Re: Solutions for a parallel system

Thanks for all who replied. If the parallel test system goes out of sync because of testing how does the subsequent application of logs work? For eg, they delete a record in the test system, at the same time the same record gets updated in the primary site. When it is time to apply the changes to standby how is the error handled. I realize that this will cause the standy to go out of sync after a while.

One of the respondents mentioned that scripts can be used to automate putting the standby in managed recovery mode and to open it. Is anyone else doing it?

Thanks

On 3/5/08, Dan Norris <dannorris_at_dannorris.com> wrote:
>
> You might consider 11g's Active Data Guard feature that allows you to use
> Physical Standby databases while using them for read-only reporting. It's
> basically the same as physical standby databases in 10g, except that you
> don't have to stop applying logs when the standby is open read-only. It does
> cost extra licensing $$$, but may be worth it depending on how important the
> requirements are to your environment.
>
> More (and prettier) info:
> http://www.oracle.com/technology/products/database/oracle11g/pdf/active-data-guard-11g-datasheet.pdf
>
> Dan
>
> Ram Raman wrote:
>
> Thanks. If dataguard physical standby works, we will consider it too. Just
> getting allthe options and deciding on the best one.
>
>
>
> On 3/5/08, Bradd Piontek <piontekdd_at_gmail.com> wrote:
> >
> > Some options:
> > 1. Dataguard - Logical Standby (not all the same restrictions). FYI:
> > 2. Oracle Streams
> > 3. Quest Shareplex
> > 4. Materialized views that fast refresh nightly
> > 5. Snap-copy (SAN vendor) of the database refreshed nightly at disk
> > level
> > 6. Transportable tablespace? (may be a shot in the dark)
> >
> > On Wed, Mar 5, 2008 at 2:30 PM, Ram Raman <veeeraman_at_gmail.com> wrote:
> >
> > > Hi all,
> > >
> > > We have been requested to come up with a solution for the following
> > > problem: We have a production system that is about 200Gb. We want to be able
> > > to build a parallel database that will resemble the production and cannot be
> > > out of sync with the production by more than 24 hrs. Dataguard was suggested
> > > as an option.
> > >
> > > With dataguard, my understanding is that the standby will have to be
> > > on a server of the same OS level, patch level, disk layout, etc. We do not
> > > have such a scenario here. Also they want to be able to work from the
> > > parallel environment (they may be making changes on their own). This means
> > > dataguard will not work as the standby will have to be in managed recovery
> > > mode. Even if we put it in OPEN mode and work on it, there is going to be a
> > > huge backlog of logs to be applied later on. I would think that would cause
> > > a big delay in refreshing? I was checking the redo log generated in prod for
> > > the past month. The average per day seems to be 20GB and maximum of
> > > ~60GB/day and min of 5G/day. Not sure how long it would to take to apply
> > > 60GB. Also this kind of setup would require manual intervention from the
> > > DBA everyday to put in managed recovery mode, open etc.
> > >
> > > Can the listers suggest options available for the problem. 10g. We
> > > plan on using this in other systems too, which are much smaller.
> > >
> > > Thanks.
> > >
> > >
> >
> >
>

--
http://www.freelists.org/webpage/oracle-l