We actually write our audit logs to a file using Log4j then we use Splunk to pull it all together. All the data in Splunk is restricted to the IdM group (though are considering exposing aggregated statistics to others). We do this because our database isn't really set up for HA (its pretty close and the likelihood of an issue is pretty minimal).
Cheers, Scott On Mon, Feb 22, 2010 at 1:28 PM, Marvin Addison <[email protected]>wrote: > On Mon, Feb 22, 2010 at 12:24 PM, Patrick Berry <[email protected]> wrote: > > I'm curious about what people are using as their data source. > > Oracle > > > Is there a performance hit if the datasource is unavailable? > > There should not be any measurable performance degradation if your > data source is unavailable, but I would encourage you to attempt to > verify those claims. My comment is based on the software design, not > on any measurement I've done. Database writes are performed by a > worker thread pool using a single thread with an unbound queue, so the > Tomcat connector thread pool handling CAS authentication would not > block or be otherwise affected. If a write fails, you might incur > some additional IO to log the stack trace to a file, but I would > imagine this would not cause a measurable degradation in performance. > > M > > -- > You are currently subscribed to [email protected] as: > [email protected] > To unsubscribe, change settings or access archives, see > http://www.ja-sig.org/wiki/display/JSG/cas-user > -- You are currently subscribed to [email protected] as: [email protected] To unsubscribe, change settings or access archives, see http://www.ja-sig.org/wiki/display/JSG/cas-user
