Re: Update schema while new data is coming in

I think that you are incorrect.
In RAC the upgrade process alter the database structure and this affect both instances.
For example:
You have a table with two columns and you do: select * into a , b from table where whatever.
Then you add a column to the table.
I think that your select will return error.

I think that this requirement needs to be implemented from the design stage. What will happen when you:

1) Add new tables
2) Add new columns
3) Increase or decrease column length
4) Change the type of a column
5) Drop a column

I am not sure that even two servers (not RAC) approach, with whatever replication method you use, will work as you will have at some point two different structures.

Adar Yechiel
Rechovot, Israel

Don Seiler wrote:
> Being able to perform updates on one member of the
> cluster at a time while the others handle the application traffic?
>
> Don.
>
> On Nov 12, 2007 3:47 AM, <Antje.Stejskal_at_ppimedia.de> wrote:
>
>> Hi,
>>
>> we are having a web application that ist o be online availyble 7x24. We plan
>> for the scenario that the application needs a software update (new tables ,
>> modified tables et. al.) Management and customer requires that the
>> application is still running while the upgrade goes on. How would you solve
>> this, that you have one server with the old schema where new data is coming
>> in, meanwhile another server is brought up to latest release. And data of
>> the old server (or delta of data) will be send to new server. We will
>> propably have 2 servers anyway because of the traffic.
>> Please help me with your ideas.
>>
>> Regards
>> Antje Stejskal
>> --
>> http://www.freelists.org/webpage/oracle-l
>>
>>
>>
>>
>
>
>
>

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