> -----Original Message----- > From: Vanga Palli, Ravindra Kumar [mailto:ravindra.vangapa...@hp.com] > Sent: Wednesday, May 22, 2013 6:46 PM > To: Tomcat Users List > Subject: RE: Need AccessLogValve results clarification > > >>>>>Note: I don't think the IP is PAT/NAT. > It depends on how you log IP, if you used %h this will be a NAT address > but if you use X-forward-for header then it is the actual client. > > Thanks, > Ravi > > -----Original Message----- > From: Jeffrey Janner [mailto:jeffrey.jan...@polydyne.com] > Sent: Wednesday, May 22, 2013 3:31 PM > To: 'Tomcat Users List' > Subject: RE: Need AccessLogValve results clarification > > > -----Original Message----- > > From: Mark Thomas [mailto:ma...@apache.org] > > Sent: Wednesday, May 22, 2013 5:24 PM > > To: Tomcat Users List > > Subject: Re: Need AccessLogValve results clarification > > > > On 22/05/2013 23:13, Vanga Palli, Ravindra Kumar wrote: > > > The timer starts when the incoming socket is assigned to a thread. > > So, what you see in logs is a processing time but not queuing time. > > > > Not quite. The timer starts once the request line has been processed > > > > Mark > > > > > > > > > > Thanks, > > > Ravi > > Thanks guys. I thought that was how it worked, but glad for the > clarification. > Now, if only I can figure out why the log shows the same request for > the login.jsp from the same IP 8 times in the same second (all with a > very long processing time). > Note: I don't think the IP is PAT/NAT. > >
Thanks Ravi, but I really meant the connection should be coming in without PAT/NAT based on what I know of the network topology (one of our client's networks). Plus, the recorded IP is in the 10.x.x.x subnet, so it's really doubtful, but I'm not totally ruling it out. If I remember correctly, the time in the log is the time when the entry was written, that is, when the request is completed. So it's possible the end-user connected once, then refreshed the page a number of times because it wasn't loading, resulting in multiple requests, all of which completed at the same time. Jeff You know, if it wasn't for end users, all software would run perfectly with great performance. --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
