Re: to many switch logfile

The recent re-implementation of log_checkpoint_timeout more or less eliminates the issue of 'long recovery times' being associated with large log files - the timeout now guarantees (nearly) that the amount of log to be applied in the event of a crash is the amount that would have been generated in 'log_timeout_seconds'.

By the way, another reason for slow performance here is that log_checkpoint_timeout has been set to 10 seconds, so Oracle is working very hard to get dirty blocks written to disk very rapidly anyway - but your estimate of 500MB for the log files looks pretty good to me.

--
Jonathan Lewis
http://www.jlcomp.demon.co.uk

Host to The Co-Operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html

Author of:
Practical Oracle 8i: Building Efficient Databases

Screen saver or Life saver: http://www.ud.com
Use spare CPU to assist in cancer research.

Sean Fitzgerald wrote in message ...


>make your redo logs bigger -- if you do lots of loading or other

>transactions, think about making them 500M or more

>

>just don't make them so big that they switch too infrequently, then you

will


>have longer recovery times if   ( a general rule is logs switches every 30

>minutes)

>

>otherwise, make the tables "nologging" and use the +APPEND hint on your

>insert statements (or use direct path sqlldr) -- this disables redo

>generation for those transactions

>

>you don't give any version/platform information or any real "detail" on how

>you are doing the inserts so this is the best I can offer

>"bartos" <bcharlin_at_elka.pw.edu.pl> wrote in message

>news:9tjaj0$d8a$1_at_news.tpi.pl...

>> Hi

>>

>>     I'm during loading data into 4 tables by insert with logging. This

>> operation has to insert about 4 milin rows to one of four tables.

>> The problem is that it is running very slow. I noticed that logfile is

>> switched every 30 sec. aproximately.

>> Additional informations:

>>

>> log_checkpoint_timeout=10

>> size of redo log 16M

>> log_buffer  1064960

>>

>> What is wrong with this.

>>

>>

>

>