Re: redo log problem

Sybrand Bakker wrote:

> Hi,
> What happens is ALL three online redo log files get written at the same
> time(ie a 'burst' of data), a fourth needs to be allocated. This means the
> database will hang until this situation has been resolved.
> Whenever a checkpoint occurs, the db needs to write out all dirty SGA
> buffers to the redo log. The message means at the time of the allocation
> error that has not been completed yet.
> You need either to increase the size of your redo log files or add a fourth
> group.

I've seen the situation which was described by the original poster and your answer brings up another question for me.

If this type of situation only occurs at peak times, and if no data will be lost, is there any real harm being done? As you describe it I can see why the database might appear to hang, but apparently all it is doing is trying to get caught up and once it is caught up operations continue as usual. I suppose that there is no harm in allocating more redo logs and if that can cut down on the amount of time that the database is in a "hung" condition during large-scale batch processing, but if that peak activity accounts for only a couple hours each night when there is no user activity (i.e. all of the client apps are logged off and only host-based batch jobs are running) then I wonder if I'd really see an appreciable decrease in the amount of time that it takes for the jobs to run?

Ken Received on Thu Jul 15 1999 - 07:14:39 CDT