Description of problem: The symptoms of the problem were, in general, sluggish network performance: 1. Slow yum updates/installs 2. "Page Not Found" errors on common sites such as google 3. excruciatingly slow smtp and pop3 mail server interaction I'm not certain, but I suspect that the networking configuration files in /etc/sysconfig/networking/... weren't properly configured with regards to ipv6, which should default to off I believe. Since this is a fairly fresh installation of Fedora 8 (about two weeks old), I suspect it was a problem either in the installation itself or in one of the subsequent yum updates. The installation is Fedora Respin 20080204. Simply going to system-config-network app, selecting the active profile under the Devices tab, and pressing the Edit button seemed to restore the misconfigured parameters and the network immediately began to function normally. Version-Release number of selected component (if applicable): n/a How reproducible: Don't know - didn't "reinstall" Steps to Reproduce: 1. Install Fedora Respin 20080204 2. Check a few well-known web sites (under Firefox) for Page Not Founds or otherwise sluggish behavior. 3. Attempt to connect to a pop3 or other mail server and obvserve behavior. Actual results: Expected results: Additional info: Linux localhost.localdomain 2.6.24.3-12.fc8 #1 SMP Tue Feb 26 14:21:30 EST 2008 x86_64 x86_64 x86_64 GNU/Linux
Note that the checkbox for "Enable IPv6 configuration for this interface" was unchecked even the first time I edited the profile. I'm guessing that the code rewrites the configuration parameters when the (profile) "Edit" dialog is exited.
well, I think you can turn on/off IPv6 in anaconda (the installer). Maybe you forgot to turn off IPv6. So this may be a bug of anaconda, not s-c-network (which fixed it)
ok Harold, i'm sorry but this whole thing may have been a farce. i found out some time after creating this bug report that there was a severe problem with roadrunner's dns nameservers in the Raleigh, NC area. After switching to 4.2.2.4 and 4.2.2.5, the problem TOTALLY went away. it may have been that doing the ipv6 thing did really help some (i.e., that this was a dual problem of both an ipv6 issue AND the roadrunner dns's), or it may have been that the perceived improvement (after ipv6 changes) was really false. sorry for being so ambiguous, but i believe being truthful makes it easier for everyone.
It's possible that they were more broken with giving back ipv6 response codes. That's happened with TWC in the past.