<DIV>Sorry for the delay David,</DIV><DIV> </DIV><DIV>The primary problem with the yukon I have seen was just plain goofy inconsistant connectivity. Duplex apparantly changing, intermittant loss of connectivity (ie 10 ping responses ok, 1 no reply, 20 ok, 1 bad....) I had tried different configurations of the switch port (cisco 3550) and different switch ports and different cables. Did not try a different machine with this exact setup, but have had similar problems with linux (fc4) on an identical machine. Beside the inconsistancies on the FC4 machine, it would throw an error on shutdown that the eth0 mac address had changed.</DIV><DIV> </DIV><DIV>The Intel nic (intel pro 1000 xt server board) has been acting well at 100Mb/s and will soon be moved to a 1G segment once I get done testing it (vmware host testing at this point).</DIV><DIV> </DIV><DIV>Take care</DIV><DIV>Don</DIV><DIV> </DIV><DIV> <DIV> </DIV>-----"David A. Ranch" <dranch@trinnet.net> wrote: ----- <blockquote style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">To: dbossung@southeast.edu From: "David A. Ranch" <dranch@trinnet.net> Date: 08/05/2006 01:52PM Subject: Re: [rhn-users] vsftpd performance very slow Hello Don, In theory, the Yukon chips were pretty nice on paper until people started using them. The Yukon chips had LOTS of bugs and though Yukon2 was better, it still has some performance related issues. Some of those issues were fixed in respins but considering their track record, are their any other issues? How's the Intel card working so far? --David dbossung@southeast.edu wrote: > David, > > I think I am experiencing some of those 'know issues' you speak of. After > emailing Daryl, the nic has oscelleted between FD and HD on a very random > basis. Just all of a sudden, for no reason, the duplex will change. > > I have already taken your advice and switched nics. All is well with the > new Intel. > > Damn Micron for using a motherboard with this chipset, been nothing but > trouble since we got the clientpros. :( > > Thanks for your input. > > Don > > > > > "David A. Ranch" > <dranch@trinnet.n > et> To > Sent by: Red Hat Network Users List > rhn-users-bounces <rhn-users@redhat.com> > @redhat.com cc > > Subject > 08/04/2006 01:09 Re: [rhn-users] vsftpd performance > AM very slow > > > Please respond to > Red Hat Network > Users List > <rhn-users@redhat > .com> > > > > > > > > It's worth noting that there are known issues with early generation > Yukon MAC and PHY chips including auto-negotiation. Check with your NIC > vendor for details or try a different NIC (Intel, etc.) > > --David > > > dbossung@southeast.edu wrote: >> Right on the money Daryl, thanks. Now getting over 10Mb/s. Much more fun >> moving ISOs around now :) >> >> FWIW I always set the switch port 100FD and then force the host to same. >> For some reason when I set this host to 100 FD with ethtool, I would > loose >> connectivity and when I set it back to HD I got connectivity. (Remember, >> the switch port is manually set to 100 FD) >> >> I set the switch port to auto and set the nic to auto, I get 100FD. Go >> figure. >> >> The switch is a Cisco Catalyst 3550 and the nic is a built in Marvel > Yukon. >> Uh, oh, That just jogged my memory that there were some problems with the >> Yukon driver under Fedora, maybe I am running into the same thing here. >> Will check and see. >> >> Thanks for the pointer >> >> >> > >> Daryl Field > >> <djf@pipex.net> > >> Sent by: > To >> rhn-users-bounces Red Hat Network Users List > >> @redhat.com <rhn-users@redhat.com> > > cc > >> 08/02/2006 05:38 > Subject >> AM Re: [rhn-users] vsftpd performance > >> very slow > > >> Please respond to > >> Red Hat Network > >> Users List > >> <rhn-users@redhat > >> .com> > > > >> >> >> >> Try copying the file using another method - scp/nfs/samba - you can then >> tell whether the network is the issue, which is most likely. >> Check IPtables isn't using 'Limit' to throttle connections - unlikely >> but I have seen this before. >> >> I'd check your duplex settings up front through. >> >> Daryl >> >> dbossung@southeast.edu wrote: >>> I have just installed RHES 4.3 and enabled/started vsftpd. >>> >>> Both the client and the server are on the same private subnet. 100Mb/s >>> >>> When I use FileZilla on a windows box to connect (using the quick > connect >>> option) and send a file (from 22M to 560M in size) the throughput is >>> extremly slow. (<100KB/s). I see similar results using WSFTP on the same >>> windows host. >>> >>> I found a kb article about using 'noreservation' in fstab, tried that > but >>> was no help. >>> >>> Any other ideas about what to look at. I don't even know how to start >> with >>> this. >>> >>> Thanks >>> Don >>> >>> >>> _______________________________________________ >>> rhn-users mailing list >>> rhn-users@redhat.com >>> <A href="https://www.redhat.com/mailman/listinfo/rhn-users" target=blank >https://www.redhat.com/mailman/listinfo/rhn-users</A> >>> >> _______________________________________________ >> rhn-users mailing list >> rhn-users@redhat.com >> <A href="https://www.redhat.com/mailman/listinfo/rhn-users" target=blank >https://www.redhat.com/mailman/listinfo/rhn-users</A> >> >> >> _______________________________________________ >> rhn-users mailing list >> rhn-users@redhat.com >> <A href="https://www.redhat.com/mailman/listinfo/rhn-users" target=blank >https://www.redhat.com/mailman/listinfo/rhn-users</A> > > _______________________________________________ > rhn-users mailing list > rhn-users@redhat.com > <A href="https://www.redhat.com/mailman/listinfo/rhn-users" target=blank >https://www.redhat.com/mailman/listinfo/rhn-users</A> > > > _______________________________________________ > rhn-users mailing list > rhn-users@redhat.com > <A href="https://www.redhat.com/mailman/listinfo/rhn-users" target=blank >https://www.redhat.com/mailman/listinfo/rhn-users</A> </blockquote> </DIV>