Tom,

I am glad I got to ask this question because some of the things became more clear reading about it, and a few more things became clearer as I was getting ready to frame my next question.

As you know, the 10g logical standby allows READ ONLY access for reporting and stuff, but with quite a few limitations imposed. But if the database were to failover to the standby side, all but the last few redo records in transit may have already been applied, so there is not much delay in recovery, and Failover. But the keyword here was limitations.

But my main confusion was about what you said "You can open the physical standby read/write ..." Did you mean to say the database open mode = "READ/WRITE" as opposed to READ_ONLY?

Questions:
1) Do I have the option to open the database READ_ONLY?
2) But if I want to do update, and open the database in READ_WRITE what is the recovery effort involved in case of a real Failover situation? Lets say I change a few records in the database, and then all of a sudden the Primary fails, then I have to unchange all the changed records and re-apply in the meantime what has changed on the primary?


Thanks,
Khalid

Khalid

In 11g you have the ability to utilize the standby as a complete standby.

You can open that standby (read-write), you can open it read-only. You can open the standby to read-write and then return to read-only. You can backup the standby and use that backup to recover the primary.

However, what you can do will depend upon whether you are using a physical or a logical standby.

Because of teh extended benefits of Dataguard in 11g, there is a license cost.

But, 11g is worth it. IMHO.

Wayne

Thanks.

Hi Tom,

I have a 2 node RAC along with singlenode standby running on oracle 10.2.0.4.

Now I am in the process of upgrading it to 11.1.0.6.
Is there any specific method to enable to active standby feature while upgrading from 10g?

Or normal Upgradation of Clusterware,ASM,Oracle Binary,Listener will workout?

Thanks.

Hi Tom,

Just want to know if we can configure 11G database with 10G logical standby ? I tried to do the same but I'm facing some issues.

Thanks

how to achieve zero downtime data migration for upgrade from 10G to 11G. i don't want to use dataguard or editions.

From the link in your response to the previous post, you suggested to create a logical standby,upgrade the logical standby and then do the switch over.

http://asktom.oracle.com/pls/apex/f?p=100:11:0::::P11_QUESTION_ID:18064818873846#3145155100346004071

Is'nt this approach going to be limited to only those databases which abide to the limitations of the logical standby?


Is it possible to have a 11gR2 Active Dataguard(Physical Standby) for a 10gR2 Primary Database?
If YES, then we could as well create a 11gR2 Active Dataguard, possibly with a Real-time apply, recover it and do the switch over. This could be an ideal way to upgrade with the least downtime possible. What do you think?

Hi Tom, I am upgrading from 10g r2 to 11g r2. I have physical standby using dataguard. While I am upgrading I will not be using any of the active dataguard features like open read write or open read only for neither Reporting nor testing purposes. This is purely for DR solution. Can I achieve the same without licensing active dataguard? If so how?

Tom,

At the top you wrote:

"o Use physical standby for testing without compromising protection of the production system. You can open the physical standby read/write - do some destructive things in it (drop tables, change data, whatever - run a test - perhaps with real application testing). While this is happening, redo is still streaming from production, if production fails - you are covered."

Can I not do this without active dg? What would I be able to do in this scenario with out it?

Also, when you use the physical standby, doesn't it ship those changes back to the primary unless you shut that off?

Tom,

I need your input on a decision I need to make for a new environment I am working on designing with an application team. Right now they are using bidirectional streams and I hate it almost more than anything. I have narrowed down their requirements to be

1) DR/replication to a failover site
2) The failover site must be open for writes for customers in the region of the country

I'm thinking logical data guard even though I have zero experience with it.

What I am trying to figure out, are the growing pains of learning this new, fairly complicated technology enough to outweigh the little knowledge I've accumulated supporting streams and the horribly complicated pains of replication.

Why does that mean I have to use bidirectional? I'm pretty sure the customers in that region don't need DR and they'll have site specific records (something the developers built into their primary keys).