We attempted to move our production web servers (config 1: apache + mod_wsgi + python + django config 2: nginx ) from 7.10 64bit server to 8.04 64bit server. Capacity tests on identical machines in every respect except 7.10 vs 8.04 resulted in a 30-40% loss of capacity in terms of number of requests per second. Problem was reproducible on multiple types of hardware.
At the time (a few months ago) we did a search and found a reported defect related to an Ubuntu & Linux Kernel patch that seemed to explain the situation. I think it was on launchpad but I've been unable to find it since. Reportedly was a confirmed issue with a relatively high priority to fix. One of our staff thought it might be related to a glibc change unique to Ubuntu but cannot recall what let to that conclusion or what the change was. Could be a red herring. Odd that we were able to discover this defect report so easily back then but are at a total loss now.
We have elected not to upgrade our servers until the issue was resolved but now 7.10 is EOL and it would appear that 9 still has this problem. Hard to imagine such a significant impact isn't more widely reported but this is where we find ourselves. Appreciate anyone who can shed some light on the issue.
Thank you for taking the time to report this bug and helping to make Ubuntu better.
Is this bug reproducible with the latest Lucid packages ?
Thanks in advance.