Description of problem: Bringing up the bugzilla logon page may occur within 2-3 seconds, but I can't begin to login until a further ~10 seconds have passed. Version-Release number of selected component (if applicable): 2.18 (Bugzilla Version 2.18-rh) How reproducible: Everytime. Steps to Reproduce: 1. go to a bug while not logged in https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=198867 2. stop other network processes and start ethereal capture all 3. click Log In in the top right corner. Actual results: Log in page is retrieved and drawn immediately within 2-3 seconds, with the mouse cursor still spinning. ~10 seconds later, the cursor changes to standard mouse pointer as the page completes and web browser inserts it's password auto-completion. Expected results: Log in page is retrieved and drawn immediately. The mouse cursor then immediately changes to show page complete and page is ready for user entry (or auto as the case may be). Additional info: firefox-1.5.0.4-1.2.fc5 on fc5+updates {2006-07-15} From Australia, round trip delay seems to be 0.3 seconds. https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=198867 Then click Log in ethereal trace shows: {35 packets} first run | repeat gw/dns need re-request|doing 00.00 -> 00.00| 00.00 -> 00.27|dns/gw lookup 00.00 -> 00.27| 00.54 -> 00.55|tcp syn 00.56 -> 00.57| 00.83 -> 01.45|encrypted handshake {second run 4 round trips} 00.95 -> 01.95| 01.86 -> 02.53|app data > page appears to be fully drawn - but mouse icon still spinning !! no traffic for 9.37| 10.32 seconds !! 11.27 -> 11.28| 12.85 -> 12.86|application data 11.28 -> 11.29| 12.86 -> 13.14|tcp reset. mouse cursor stops spinning and firefox password fill puts my username/password into the boxes. The last packet before the ten second delay is from me to bugzilla (https ACK). The next message is from bugzilla to me (with application data, but very small packets that seem to be finalizing the connection).
Created attachment 132519 [details] 2nd packet capture for retreiving the login page ethereal libpcap format
*** This bug has been marked as a duplicate of 214842 ***