How to ensure that Data Guard is reliable? [message #420414] |
Sat, 29 August 2009 05:38 ![Go to next message Go to next message](/forum/theme/orafaq/images/down.png) |
Smith_X
Messages: 56 Registered: January 2007
|
Member |
|
|
Hello,
I have production system and it was set to transport archived redo log to backup system at DRP site with Oracle Data Guard.
The question is, how do I ensure that all data are transfer completely and accuracy from production system to DRP site. Are there anything I can check? I try to check the exact data in each table but that will take a 100 of years.
Any recommendation please?
|
|
|
|
|
|
Re: How to ensure that Data Guard is reliable? [message #424546 is a reply to message #420414] |
Sat, 03 October 2009 05:47 ![Go to previous message Go to previous message](/forum/theme/orafaq/images/up.png) ![Go to next message Go to next message](/forum/theme/orafaq/images/down.png) |
Smith_X
Messages: 56 Registered: January 2007
|
Member |
|
|
Hello,
I think I found something from view v&xxx and sp file. (I read from Data Guard Administration and Concept book. It's very handy.
However, I have question again and it seems I can't get answer from the Data Guard book or Google.
I want to know the value of archive_lag_target. What's the different between set it to Zero (0) or set it to a specific number?
The manual recommend typical number as 30 minutes (1800) but default value is 0. So, 0 means Log Switch will automatic perform in real time?? or it won't do anything at all ??
which value is better? 0 or 1800 (30 minute) please?
|
|
|
|
Re: How to ensure that Data Guard is reliable? [message #424560 is a reply to message #424555] |
Sat, 03 October 2009 15:03 ![Go to previous message Go to previous message](/forum/theme/orafaq/images/up.png) ![Go to next message Go to next message](/forum/theme/orafaq/images/down.png) |
Smith_X
Messages: 56 Registered: January 2007
|
Member |
|
|
Thank you very much Michel Cadot,
I saw that sentence you quoted for me. However, I'm not understand. Could you please a little bit explain to me.
I just want to confirm what I understand is correct.
- 0 is turn it off (log transport never pump log to standby database) or log transport will perform in real time ??
- 30 is optimum value
and which value I should set please?
Because in manual said
30 is typical value / optimum value
0 is default value
I do not know which one is better T_T
|
|
|
|
Re: How to ensure that Data Guard is reliable? [message #424579 is a reply to message #420414] |
Sat, 03 October 2009 23:17 ![Go to previous message Go to previous message](/forum/theme/orafaq/images/up.png) ![Go to next message Go to next message](/forum/theme/orafaq/images/down.png) |
Smith_X
Messages: 56 Registered: January 2007
|
Member |
|
|
Ok. I understand that better is hard to tell.
But could anyone please confirm me my understand is correct?
30 -- Log Writer pump log to Standby Database every 30 minute.
0 -- turn off
What does turn off mean?
1. Log Writer never pump log to Standby Database. I or DBA need to do manual recover the database by use RMAN. or
2. Log Writer automatic pump log to Standby Database in realtime. The disadvantage is.. severe downgrade the whole performance but the availability of the Standby Database is very most update (in nearly real time)
I think (2) is the answer but I'm not sure.
|
|
|
Re: How to ensure that Data Guard is reliable? [message #424580 is a reply to message #424579] |
Sat, 03 October 2009 23:31 ![Go to previous message Go to previous message](/forum/theme/orafaq/images/up.png) |
![](/forum/images/custom_avatars/136107.jpg) |
BlackSwan
Messages: 26766 Registered: January 2009 Location: SoCal
|
Senior Member |
|
|
ARCHIVE_LAG_TARGET=0
When lag is 0 , then changes are immediately (0 delay/lag) sent to standby.
ARCHIVE_LAG_TARGET=30
If/when a "bad" DML occurs on primary, you have 30 minutes to prevent it from impacting standby.
When primary has fatal meltdown, then LAG=0 is better.
When wrong DML occurs, then LAG=30 is better.
Which type of failure do you need the insurance of the standby?
There is NO correct answer.
It must be decided for each DB, on a case by case basis.
|
|
|