-
Gurus,
In order to have replicated schema (or the entire database) froma production database, can Replication and using LogMiner be compared ? On the face of it, both strategies seem capable of delivering expected outcome. But, which one should be the choice ?
Compare them based on following aspects (and of course any additional ones that you deem necessary) :
01. Technical ease
02. Overheads added to production database
03. Overheads added to production database server
04. Feasibility of starting with a restored hot backup
05. Simplicity of recovering in case the thing breaks
Additional information :
OS : HPUX 11 (64 bit)
Database : 8.0.6 (32 bit)
Db size : 200G (Archivelog)
Avg daily addition : 500M
# of archivelogs per day : @50 (100M each)
One refresh of the mirror database per day is enough.
Thanks in advance to all of you and expecting some good discussion...
-svk
-
Why dont you use standby database? definitely NOT logminer, standby is like a cheap version of replication based on archived LOGS
-
Because I am not looking at this as a solution for fast recoverability or HA. I need this "replicated" database to be opened for reporting needs. So obviously, need to be able to create some intermediate tables and a few indexes that do not exist in production db.
-svk
-
Well then you cannot use Replication neither, if you want to use LOGMINER to syncronize two database I wish you luck! How are you gonna apply changes in one log manually? If you generate 500MB Logs per day youprobably spend 2 or3 days applying those changes manually no?
-
Why cann't I use replication for reporting ? Is it that I can not open the replicated database for use ? Are there any restrictions on this like those for a standby (as in 8.0.6) ?
And applying those logs should not take days. I could create a script that syncs the databases based on previous night's SCN related to hot backup. It works. So, will not be doing it manually.
-svk
-
Am I getting Confused here? Replication using logminer???? why would you want to use the logminer in replication, when you can do a snapshot refresh/ read-only snapshots. Do you need the doc for the replication? If so, pl. let me know.
Sam
-
What I am basically looking for is a way to have a database that is at the most a day behind the production db. I need this db to be open always and be able to create a few indexes and create some intermediate tables in it. All this for the "reporting" requirements of the application.
Now, I have two "options" to establish a such a "reporting" database. One is doing it through ORACLE's replication manager (snapshot based) and the other is using LogMiner for generating the REDO statements and applying them on this "reporting" database. Either way, the "refresh" or "REDO application" process will run once a day.
So, from this point of view, I want to compare the two approaches. Hope, this is more clear than my earlier posts.
And of course it will be great if you can gice me some documentation on replication. Especially, if it talks about setting one up using a restored/recovered hot backup. Most of the documents I have come across talk about doing this using exp/imp.
Thanks and expecting more from all of you.
-svk
-
I would pesume that replication would be the right way to go. Where there are things that are already there. You can look in the logminer and its a tedious approach as far as I'm concerned. I hope you would have come across this doc, if not take a look into it on how to setting up the replication. I'm dono whether it talks about exp/imp or restore&recover
http://www.oradoc.com/ora8doc/DOC/se...651_01/toc.htm
http://www.oradoc.com/ora8doc/DOC/se...0_osr.htm#2299
Sam
-
Thanks for the links Sambhvan !!
