This content has been marked as final. Show 4 replies
You can find some Data Guard advantages in High Availability Overview relevant section http://docs.oracle.com/cd/E11882_01/server.112/e17157/architectures.htm#i1008366.
But first you should define your high availabiliity requirements especially RTO and RPO http://docs.oracle.com/cd/E11882_01/server.112/e17157/hadesign.htm#i1005920 .
Hello;1 person found this helpful
Here's some questions I be asking :
If you have an issue with the SAN solution will Oracle support bail you out?
Has Oracle endorsed this SAN Solution in any shape or form?
How do you know whether your SAN DR solution actually works without doing a failover?
Can your company tolerate 10 mins of transactional data loss?
What additional (productive) benefit does your SAN DR deliver?
Your database volumes are spread across a number of LUNs on the SAN. ALL the LUNs must be of the same point in time, consistent with each other. How can the SAN do that? ( even if it can you break a key rule of Oracle recovery - Never recover more than you have to )
The SAN solution will probably fall short because it does not understand transactions.
What it comes down to is second guessing how Oracle recovery works and thinking that some 3rd part solution is better.
None of the other technologies do what Data Guard does so it is a bit hard to compare. Sort of like asking what are the advantages of the Oracle Databases versus SHA1 and AES256.
Only Data Guard, of the technologies you've listed, will check everything it is replicating to make sure there is no physical or logical corruption. So if you don't care whether your standby will actually stand-up when required any technology will do. But if you want your Standby to be there for you when you need it ... Data Guard is the only choice available.
Should you care whether your standby is physically and logically perfect is sort of like asking whether you want the fire extinguisher at the end of the hall to work. You may never need it but you are going to be really unhappy if you need it and it isn't there. Remember the faster the replication, for example SAN-to-SAN, the less likely it will be looking at the logic of what it is replicating. SAN-to-SAN will just as happily replicate "00000000000000000000000" as it will a valid block.
1 person found this helpful
user8983130 wrote:You posted DG vs Metro. It's impossible answer what is the better for you, since you don't posted your intention... what's your goal?
If any organization wants to go for DR solution, which option would be better....DG or Metro Cluster(SAN to SAN replica)??
Metro cluster is a expensive solution , while there is a good option for data guard.(my opinion)
Please give some adv. and disadv. of both so that i can choose one for the next 5 years....Cheers!!!
The correct answer is: Choose the technology that achieve your goal. Metro and DG technologies are completely different, though "seem" that does the same thing (ie replicate).
Without being afraid of making a mistake. I would choose DG because it's a technology designed specifically for database and has many features that other technologies does not have.
Metro - I like to use the Metro to migration, backup or data replication for general purpose.
* We don't need a Host for that purpouse
* We can make sure that all data will be replicated.
DG - I like to use DG when matter is HA of Database
* It ensures the integrity of the database, does not replicate physical or logical errors.
* We can backup a STBY database
* We can open this database in read only to balance the load between two env. (i.e It's not a dead Database while replicating)
* We can open the STBY database in read/write mode to test purposes (homologation of applications) or export (using EXPDP) using SNAPSHOT feature. (11.2)
* We can use this database to UPGRADE version with minimun downtime.
.... and so on.
See about MAA:
See Studies Case of DG:
I may be sinning, because I had bad experiences with replication in the level of storage (peer-to-peer) for the database stored on it.
Since the database is corrupted on the target it will be replicated to destination, then will be corrupted in both sites.
Have you ever imagined spending a fortune on technology/hardware and when you need it, it does not meet to the expectations, it is a bad feeling.
My advice: Study both technologies and see what fulfills your expectation. You can call a representative pre-sales from each vendor to show the use cases.