Skip to Main Content

GoldenGate

Announcement

For appeals, questions and feedback about Oracle Forums, please email oracle-forums-moderators_us@oracle.com. Technical questions should be asked in the appropriate category. Thank you!

Need Help Understanding GoldenGate DR (Standby) Switchover/Fail-over with DBFS and Fuser

user13691955Feb 9 2015 — edited Feb 9 2015

Hi, Gurus. We are about to embark on GoldenGate DR failover / switchover configuration and testing using DBFS and Fuser. The details are described in Doc ID 1436913.1 and 1322547.1.

I know the basics of GoldenGate and Oracle DataGuard. But DBFS and Fuser are new concepts to me. I need some help.

1) Is GoldenGate installed in a DBFS partition on the primary? SO, DBFS has to be configured before it is installed?

2) Are these GoldenGate files and directories (ggsci, /dirdat, /dirrpt, etc) visible from UNIX level? Will I be able to navigate in the GG mount or is it a closed file system like ASM?

3) DBFS has to be a part of database. So, what happens if primary database crashes and cannot be mounted? My guess DBFS will not start because the database is down. Or does it run outside of database?

4) Does DBFS store just links to large files in the Oracle database, but the actual GoldenGate binaries files still reside on OS level? Ot they are INSIDE of the database - both GG binaries and trails?

5) How does GoldenGate replicates itself to standby? Because it is a part of database and primary archivelogs and redo logs send the changes to the standby? Or it ONLY runs until switchover/failover on the primary, but afterward it is simply unmounted from primary and mounted to standby by the oracle OS user and the scripts? Or is there a datapump and extract that send trails to standby and after the failure those trails are simply being used?

I guess that is my main question. Why do we bother with DBFS at all? We can simply add a standby extract/datapump to the primary. It will send trails to the standby, that will have only manager running. The manager will accept the trails, periodically deleting all older than lets say one week. At the time of failover/switchover we simply use the same queries as in the docs above to find out if standby became the new primary (the failover had occured). If it did, we can use the same script to get SCN from which this new incarnation started , use a shell script to use an obey file to add the replicat starting from that sequence and voila! Why do we need to bother with fuser and cumbersome DBFS at all? Please let me know what I am missing. If there are alternative to this DBFS DR GoldenGate switchover I would like to use which ever one is the simplest. Thanks.

Comments
Locked Post
New comments cannot be posted to this locked post.
Post Details
Locked on Mar 9 2015
Added on Feb 9 2015
0 comments
653 views