Re: 8.1.6 goes mad until HD full

"Hannes Erven" <h.e_at_gmx.at> wrote in message news:3bd89790$0$18120$6e365a64_at_newsreader02.highway.telekom.at...

>

Sybrand is right that your redo logfiles are too small, and that you shouldn't just delete them when the archived redo log device becomes full. After all, you switched on redo log archiving for a purpose - to *archive* your redo logs.

I'd recommend using logminer to view the contents of these redo logs. See my web-page:

    http://home.clara.net/dwotton/dba/logminer.htm

for information on how to do this.

I had a similar problem at a previous employer. An application had been designed so that large data uploads were performed overnight, with no periodic commit. In the morning, the user would come in, look on the screen to see if there had been any validation errors and either press the "commit" or "rollback" button. If they pressed "rollback" the database would then proceed to rollback about 8 hours overnight work in the next 20 minutes, generating huge amounts of redo and dragging the server to its knees (at peak usage time!).

Looking at the v$sqlarea might not reveal this - just a single rollback statement can generate a lot of redo.

HTH, Dave.

Dave.

--
If you reply to this newsgroup posting by email, remove the "nospam"
from my email address first.