No question at this time
DBA Top 10
1 A. Kavsek 10900
2 M. Cadot 6200
3 P. Wisse 5000
4 B. Vroman 4500
5 T. Boles 2200
6 . Lauri 2000
7 J. Schnackenberg 1300
8 J. PĂ©ran 1100
9 R. Pattyn 700
10 T. P 500
10 G. Graham 500
About
DBA-Village
The DBA-Village forum
Forum as RSS
as RSS feed
Site Statistics
Ever registered users48425
Total active users1512
Act. users last 24h8
Act. users last hour1
Registered user hits last week172
Registered user hits last month818
Go up

Regd. Bounded Recovery Checkpoint and BR Interval Regd.
Next thread: Managing shared memory caused by ORA-04031
Prev thread: row to columns

Message Score Author Date
Hi Friends, Regd. Bounded Recovery, 1. When does...... Jalal B Sep 09, 2019, 07:21
Are you talking about FAST_START_MTTR_TARGET? ...... Score: 400 PtsScore: 400 PtsScore: 400 PtsScore: 400 PtsScore: 400 Pts Michel Cadot Sep 09, 2019, 09:38
HI Michel, I think OP question is related to Go...... Score: 100 PtsScore: 100 PtsScore: 100 PtsScore: 100 PtsScore: 100 Pts Ales Kavsek Sep 09, 2019, 10:33
Ah, OK! Sorry for the confusion. Regards Mic...... Michel Cadot Sep 09, 2019, 17:39
Thanks Ales and Michel. Yes,this is related to Go...... Jalal B Sep 10, 2019, 04:57

Follow up by mail Click here


Subject: Regd. Bounded Recovery Checkpoint and BR Interval Regd.
Author: Jalal B, India
Date: Sep 09, 2019, 07:21, 11 days ago
Os info: linux
Oracle info: 12c , gg version 11
Message: Hi Friends,
Regd. Bounded Recovery,
1. When does the Bounded Recovery checkpoint triggers first?
2. IS BR Checkpoint and BR Interval triggered only after a longrunning transaction is
started?
3. Long Running Transaction as per Bounded Recovery:
Is it when a Transaction is open for 4hours and more?

4. BR interval - From when does it start. It Starts with the Start of the first Long Running Transaction?

I read documents, not able to understand.

Thanks, Jalal
Goto: Reply - Top of page 
If you think this item violates copyrights, please click here

Subject: Re: Regd. Bounded Recovery Checkpoint and BR Interval Regd.
Author: Michel Cadot, France
Date: Sep 09, 2019, 09:38, 11 days ago
Score:   Score: 400 PtsScore: 400 PtsScore: 400 PtsScore: 400 PtsScore: 400 Pts
Message:
Are you talking about FAST_START_MTTR_TARGET?
If so it is not related to long transactions.
It starts with the instance if this parameter is set.

First, a remind, the main work of instance recovery is to rebuild blocks that have been modified (dirty blocks in cache) and committed but not written onto disks before instance crash (the other part is to rollback dirty blocks that have not been committed but nevertheless written onto disks). So the main point is about number of dirty blocks and so checkpoints.

Then, a little bit of history.
Long time ago no parameter existed and you couldn't know how long an instance recovery will last after an instance or server failure. Less than a minute or hours? It depends on the activity of the instance and when the last checkpoint happened. The only way to handle an upper value of time was to limit the size of the redo logs (as a checkpoint occurred when there is a log switch).

So to overcome this Oracle introduced 2 parameters:
- LOG_CHECKPOINT_TIMEOUT to force at least a checkpoint every number of seconds specified by this parameter
- LOG_CHECKPOINT_INTERVAL to force a checkpoint when the number of redo blocks given by this parameter have been generated.
This mechanism is not the most appropriate one:
- Specifying a timeout, if the activity is low then the checkpoint is useless (but this does not hurt), if your activity is occasional high then your parameter will maybe too high to avoid many dirty blocks during this period and if you specify a very short time you increase the server load during the other periods, and for maybe nothing: for example, you modify a single block every millisecond, you will generate many redo blocks but only one buffer has to be rebuilt by the instance recovery and checkpointing it from time to time is sufficient.
The problem with this parameter is that it addresses the redo log activity when checkpoint (and then instance recovery) is based on the dirty buffers in the cache.

So Oracle introduced the FAST_START_IO_TARGET to limit the number of IOs during the instance recovery by limiting the number of dirty blocks in the cache checkpointing when this number number exceeds the value of this parameter. This was a real improvement but we, DB users, think our recovery in terms of time and not of IOs.

So FAST_START_IO_TARGET was then replaced by FAST_START_MTTR_TARGET specifying a target for the recovery time (in seconds).
Oracle has to convert this time in IOs and so in number of dirty buffers before checkpointing.
The current computed values for these points can be seen in V$INSTANCE_RECOVERY:

SQL> desc V$INSTANCE_RECOVERY

Name
--------------------------------
RECOVERY_ESTIMATED_IOS
ACTUAL_REDO_BLKS
TARGET_REDO_BLKS
LOG_FILE_SIZE_REDO_BLKS
LOG_CHKPT_TIMEOUT_REDO_BLKS
LOG_CHKPT_INTERVAL_REDO_BLKS
FAST_START_IO_TARGET_REDO_BLKS
TARGET_MTTR
ESTIMATED_MTTR
CKPT_BLOCK_WRITES
OPTIMAL_LOGFILE_SIZE
ESTD_CLUSTER_AVAILABLE_TIME
WRITES_MTTR
WRITES_LOGFILE_SIZE
WRITES_LOG_CHECKPOINT_SETTINGS
WRITES_OTHER_SETTINGS
WRITES_AUTOTUNE
WRITES_FULL_THREAD_CKPT


And, if you have bought the option, its history can be seen in DBA_HIST_INSTANCE_RECOVERY.

Regards
Michel
Your rating?: This reply is Good Excellent
Goto: Reply - Top of page 
If you think this item violates copyrights, please click here

Subject: Re: Regd. Bounded Recovery Checkpoint and BR Interval Regd.
Author: Ales Kavsek, Slovenia
Date: Sep 09, 2019, 10:33, 11 days ago
Score:   Score: 100 PtsScore: 100 PtsScore: 100 PtsScore: 100 PtsScore: 100 Pts
Message: HI Michel,

I think OP question is related to Golden Gate Bounded Recovery:

http://rob.lasonder.org/index.php?title=Golden_Gate_architecture:_GoldenGate_Bounded_Recovery

I'm afrtaid not that many of us here have hands on experience with GG....

Regards,
Ales

Your rating?: This reply is Good Excellent
Goto: Reply - Top of page 
If you think this item violates copyrights, please click here

Subject: Re: Regd. Bounded Recovery Checkpoint and BR Interval Regd.
Author: Michel Cadot, France
Date: Sep 09, 2019, 17:39, 10 days ago
Message:
Ah, OK! Sorry for the confusion.

Regards
Michel
Your rating?: This reply is Good Excellent
Goto: Reply - Top of page 
If you think this item violates copyrights, please click here

Subject: Re: Regd. Bounded Recovery Checkpoint and BR Interval Regd.
Author: Jalal B, India
Date: Sep 10, 2019, 04:57, 10 days ago
Message: Thanks Ales and Michel.
Yes,this is related to Golden Gate.

Ales,
I read the doc which you shared, and did not understand,the point, "when" the first BR Checkpoint triggers.

Thanks, Jalal
Your rating?: This reply is Good Excellent
Goto: Reply - Top of page 
If you think this item violates copyrights, please click here