02-28-2015 07:42 AM
Hi All;
we are looking for whether VVR can use for the oracle database replication instead of oracle data guard solution.
If it is used, do you know veritas gives support for any problem faced.
even VVR keeps the write order fidelity, it is not certain the database integrity will ve preserved at the disaster site.
do you have any best practices and white papers, experience, anything you suggest for this deploymeny?
03-06-2015 01:46 PM
It is now simply called Replicator Option and can certainly perform the replication instead of Oracle Data Guard. I could not find the documentation swiftly; however, it really depends upon the architecture including distances. Synchronous of course being the most flawless but limited on distance for obvious reasons.
03-06-2015 02:07 PM
The VERITAS Storage Foundation for Oracle RAC Best Practices Guide is a good source albeit a little outdated.
03-09-2015 05:24 PM
Hello Asiye,
If your are going to asynchronous replicate redo logs and database files it is not going to work. Data loss on active site might corrupt the VVR back log along with data.
Synchronous replication should work, check this Oracle article about Sun Cluster
http://docs.oracle.com/cd/E18728_01/html/821-2847/gbbsi.html
It should be applicable to VCS with VVR just as well as to Sun Cluster and SDRF/s.
Also you could use VVR for archive log shipping, keep standby database in mount mode and run rman script every X minutes(depending of amount of archive log generation) to recover the database.
On the active end you will need to do
ALTER SYSTEM SWITCH LOGFILE;
every so often.
Another part of "DG" will be IP address\hostname migration. However, you will need a standby database licensed and provided with storage.
Also you might just get away with materialized views without VVR or DG, depending on the application.
03-10-2015 07:19 AM
You can use Async VVR as long as your redo logs are in the same diskgroup as data files - see discussion https://www-secure.symantec.com/connect/forums/vvr-oracle-several-dgs-part-sf-hadr, so if redo logs are in the same diskgroup as data files then your data will remain consistent, even after a network disconnect and then a reconnect where VVR drains from SRL in write order. I don't know of any other replication technology that does this as all other replication technologies I know about do not keep the data consistent when catching up from a disconnect. The only time VVR will because inconsistent is if your SRL fills and even then it will not become inconsistent until after a manual task to fix full SRL.
I would also recommend you exclude temp DB from replication - see discussion https://www-secure.symantec.com/connect/forums/need-solution-veritas-storage-foundation-ha
Mike