Hello Ryan, On Thu, Jun 05, 2008 at 10:41:57AM -0600, Ryan Kish wrote: > > Why do you think you need udp to get nfs failover to work? > > > > I am under the impression that using TCP would cause a TIME_WAIT issue > during node failover. Its actually stated at hint #2 on the Linux-HA > website's NFS page: http://www.linux-ha.org/HaNFS
gosh, who has written this site? Especially with 2.6. kernels you *should* use tcp, since 2.6. suffers more from lost udp fragments than 2.6. did (I'm not sure if this is fixed in the mean time or will be ever fixed). Anyway, there is certainly no 15-20min timeout. Maybe it is a typo and should be 0.5min to 2min? Yes, this timeout is there, but so it is with udp. > If this is incorrect, I am more then happy to switch over to TCP! Test it yourself, if you run into a 20min NFS timeout you certainly hit a bug and should switch the kernel version ;) Cheers, Bernd _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
