Eric,
I have been reading this thread and been thinking on the same lines as
others did. One thing comes to my mind is, what happens to the application
program when you see log switching has stopped? Are there any tools to
monitor the application program's state when that happens? Does the
application get stuck in a loop and stop processing? May be there is nothing
wrong with the database but application program logic is causing this
behaviour...
- Kirti
> -----Original Message-----
> From: Eric Fang [SMTP:eric_fang_at_yahoo.com]
> Sent: Wednesday, May 03, 2000 1:29 PM
> To: Multiple recipients of list ORACLE-L
> Subject: Re: Log Switch hangup?
>
> Jared :
> Thanks for the reply. I know the log switch stops
> because 1) it doesn't advance in v$log_history for 2
> hours 2)the pplication will commit after inserting
> every 3000 records. When I query the table, it still
> has only 3000 records even after 2 hours.
> I am able to login the database and do all the
> queries. Since it's just a test database, it's not in
> archivelog mode.
> I will set log_checkpoint_timeout to 1200, how
> about the log_checkpoint_interval (I use 10000), do I
> also have to increase? The redo log file is 20m each
> (4 redo logs).
> The application basically has one user with one big
> table of 1.1m records and some other small tables. In
> the java application, it has 15 connections with 12
> threads, has some data contension when the 12 threads
> insert into another table at the same time(also 1.1m
> records).
> Glad to have the list and learn from you about the
> database tuning and application tuning stuff. Thanks.
>
> Eric Fang
>
> --- Jared Still <jkstill_at_bcbso.com> wrote:
> >
> > How do you know the log switch stops?
> >
> > Are you able to login to the database?
> >
> > It sounds like you have log archiving on, but
> > the destination directory is full. The database
> > will hang until you make room for new log files.
> >
> > Or you have the database in archivelog mode, but do
> > not have the archiver process running.
> >
> > Also, switching logs every 30 seconds is too
> > frequent.
> >
> > Resize your logs so that they never switch less than
> > 20 minutes apart, and set log_checkpoint_timeout =
> > 1200
> > in your init.ora so that checkpoints are ensured at
> > 20 minute intervals.
> >
> > Jared
> >
> > On Tue, 2 May 2000, Eric Fang wrote:
> >
> > > The database is Oracle8i on NT. After certain
> > > transactions(delete 1m records from a table), the
> > log
> > > switch stops(I checked v$log_history), usually it
> > > switches every 30 seconds, I have to tune the redo
> > log
> > > by either increase the redo log file(one user with
> > 4
> > > redo log file, 10m each) or increase the
> > > log_checkpoint_interval(I set the value as 80000
> > with
> > > db_block_size=16k).
> > >
> > > Could anyone give me some advice? Thanks.
> > >
> > >
> > > ERic Fang
> > >
> > > __________________________________________________
> > > Do You Yahoo!?
> > > Send instant messages & get email alerts with
> > Yahoo! Messenger.
> > > http://im.yahoo.com/
> > > --
> > > Author: Eric Fang
> > > INET: eric_fang_at_yahoo.com
> > >
> > > Fat City Network Services -- (858) 538-5051
> > FAX: (858) 538-5051
Received on Wed May 03 2000 - 18:06:22 CDT
