On Fri, 2 Apr 2004 10:21:23 +1000 Norman Gaywood <norm@xxxxxxxxxxxxxxxxx> wrote: > On Thu, Apr 01, 2004 at 04:31:57PM -0500, Chris Miller wrote: > > On Thu, 2004-04-01 at 09:48, Mitch Wiedemann wrote: > > > Is the FC1 2.4.22-1.2174.nptl kernel still having problems with > > > SMP and hanging? > > > > I run around 100 SMP servers with no problem. > > > > Hardware. > > Dual Xeon servers with HT on supermicro server boards. > > Dual AMD on Tyan 2469,2468,2462 > > Dual Opteron on Tyan 2880 > > > > And I do run the fedora kernel. Anything I can do to help let me > > know. But I have never seen any of the problems anyone is talking > > about. > > I'd be interested if you could run the script from: > > http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=109962#c40 > > and see how you go. The script creates a small file in /tmp, makes a > filesystem image out of that file, and then continuously does a > loop-back mount/umount until you get sick of it and ^C out. > > Any FC1 SMP kernel I have used on several types of systems, including > AMD single processor systems running the athlon SMP kernel, will > lockup in less than 500 mount/umounts. Usually much faster. > Our servers have ~10 partitions mounted, all local ext3 and have no problems with the SMP kernel. However the workstations (with hyperthreading switched on) automount some of their disks from the servers. They tend to lock up after a few days. So one workaround appears to be to use fixed nfs mounts at boot time. FWIW, running the script, the workstations would lock up between 1000 and 1400 mounts. Regards, Chris
