Greg Kuhnert wrote: > I have a server with the same chipset. amd I've been dealing with > problems ever since. This relates to a known bug in the drivers for this > card. A fix has been posted back in October last year.... > > http://userweb.kernel.org/~romieu/r8169/2.6.27-rc9/20081007/0009-r8169-fix-RxMissed-register-access.patch > > Its a shame that no-one has integrated this into the official Centos > kernel code yet. I also have a server with this problem. Rather than > running custom unsupported kernels, I have a small script that runs to > detect and recover from the problem. Basically, it pings the default > gateway. If the response time is greater than 50 msec on average, it > does a ifdown / ifup for that interface. I run this every 15 minutes in > /etc/cron.quarter-hourly
Is there a reason that running such a script is preferable to using a better-supported NIC? We've got dozens of spare NIC's in our parts closet we keep around for just this sort of occasion. There are a few boxes in the room that we just don't use onboard NIC's on at all. Just easier to disable them and use one you know will "just work." -- Chris Gebhardt VIRTBIZ Internet Services Access, Web Hosting, Colocation, Dedicated www.virtbiz.com | toll-free (866) 4 VIRTBIZ _______________________________________________ Blueonyx mailing list [email protected] http://www.blueonyx.it/mailman/listinfo/blueonyx
