Of course if you use redo transport services in the mode where lgwr sends
directly to the standby redo logs, this is not an issue.
All three of "Maximum Protection," "Maximum Availability," and "Maximum
Performance" CAN use LGWR as the redo archival process ("Maximum Proformance"
CAN use ARCH.)
mwf
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On ;
Behalf Of Mladen Gogala
Sent: Monday, October 16, 2017 10:12 PM
To: Reen, Elizabeth; oracle-l@xxxxxxxxxxxxx
Subject: Re: log file size
ARCHIVE_LAG_TARGET should take care of that. 4GB redo log saves the day only in
the situation in which there is a log switch every minute or two.
ARCHIVE_LAG_TARGET will force switch even if the redo log is not full yet.
Regards
On 10/16/2017 03:19 PM, Reen, Elizabeth wrote:
Another thing to look at is the transfer rate of your network. You
want as many logs to get over to the DG environment as possible.
Liz
Elizabeth Reen
CPB Database Group Manager
718.248.9930 (Office)
Service Now Group: CPB-ORACLE-DB-SUPPORT
-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Mladen Gogala
Sent: Monday, October 16, 2017 10:53 AM
To: oracle-l@xxxxxxxxxxxxx
Subject: Re: log file size
Hi Orlando,
There is not theoretical size limits to your redo log files. If you have
enough storage, you can make your redo log size 1PB each. However, the larger
your redo log files are, the longer it will take to apply them.
4GB is a normal redo size for OLTP databases with a lot of concurrent
transactions. I have even encountered 16GB redo logs at one large bank.
However, I would advise setting log_checkpoint_timeout to induce a checkpoint
at least once every 10 minutes. That way, you will only need to recover last
10 minutes, no matter how large your redo log files are.
Regards
PS:
----
If you opt for 1PB log files, please let me know. I would be very curious
about that.
On 10/16/2017 02:02 AM, Orlando L wrote:
Hello,
I recently inherited a prod DB; one of the things I noticed was that
it was doing log switches around 15 times a minute (not one hour). I
did some calculation based on the log sizes and I came up with a size
of 4 Gb per log for less frequent switches. Is there any limitations
fellow oraclites can think of in setting the online log files high.
One thing I can think of archiver delay locally and standby lag, if
any, in case if the switches slow down.
Orlando.
