Re: EMC storage replication

  • From: Fuad Arshad <fuadar@xxxxxxxxx>
  • To: goran00@xxxxxxxxx
  • Date: Thu, 2 Dec 2010 09:31:12 -0800 (PST)

For Point #1 
ON  A Server Crash How do you plan to have the application connect. The Good 
thing about EMC is it can replicate everything. but if the application doesnt 
have datasource built in. There are a lot of manual steps in play.
Frequent testing means of you add a disk device you need to modify EMC scripts 
to reflect the disk devices since EMC replicates and the  device level not at 
the filesystem or volume group level. A device mismatch can cause the 
replciation to stop and if proper monitoring framework is not in play will 
issues and headaches.

From: goran bogdanovic <goran00@xxxxxxxxx>
To: K Gopalakrishnan <kaygopal@xxxxxxxxx>; fuadar@xxxxxxxxx
Cc: ORACLE-L <oracle-l@xxxxxxxxxxxxx>
Sent: Thu, December 2, 2010 10:54:14 AM
Subject: Re: EMC storage replication

Gopal, Faud,

Thanks for inputs ... I appreciate a lot.
I am aware of the fact that SAN replication on block level is not aware of 
'content' of Oracle datablocks i.e. if there is a corruption on primary the EMC 
Replication will happily replicate it to secondary SAN.
With this I can live. The reason is following: we plan to use it for 
not-so-critical DB-systems - for critical ones we have DataGuard ;-)

If the corruption happens and being detected within DB backup retention policy, 
we can always restore to state before corruption and re-load the rest of the 
data - business logic allows this. Since we already have EMC Replication 
license, I see no reason to spend money on additional DG license for such 
But here is the key-point to detect corruption within DB Backup retention 
... if not, hmmm ... I have to think about this.

What makes me worry, are the points from Faud.
1. What do you mean when you say "Application Architecture needs to be there"? 
What exactly? TAF? FAN? or similar?
2. "Frequent testing of this infrastructure needs to be there" - would you 
please specify more this point?

Best Regards,

On Thu, Dec 2, 2010 at 5:11 PM, K Gopalakrishnan <kaygopal@xxxxxxxxx> wrote:

>You got great inputs so far. Just a word of caution, the EMC replication is at 
>the array level and it has no knowledge about Oracle datablocks.  If there is 
>corruption (believe me, it happens) at the oracle datablock level, storage 
>will not aware of that and propagates the corruption to the secondary (or 
>remote) SAN. You have to periodically validate the readiness of the DR site 
>SAN mirroring is used.
>You don't need to do this additional step when DG is used as it validates 
>during transfer and recovery..

Other related posts: