On Mon, 12 Jan 2004, Don Lewis wrote: > >> On 12 Jan, Mikhail Teterin wrote: > >> > >> > . 5.2-CURRENT (Dec 14) server, RedHat-9 client: > >> > core is created properly, but sometimes the server goes > >> > into a frenzy with the sys-component (bufdaemon) taking > >> > up the entire 100% of the CPU-time (P4 at 2GHz); it only > >> > writes _at_4Mb/s (~14% of the disk's bandwidth) and the > >> > only cure is to restart the /etc/rc.d/nfsd; trying to, > >> > for example, switch from X11 to a textual console, when > >> > this is happening reliably hangs the machine. > >> > >> I saw something similar a in the last month when running iozone on a > >> FreeBSD client with an nfs file system mounted from a FreeBSD server. I > >> think this was in the 5.2-BETA timeframe. The client typically ran out > >> of CPU first, but the server was not far behind. This happened in the > >> tests with the larger file and/or block sizes. Bufdaemon was typically > >> the big consumer of CPU. > > > > Hm, I can attempt to reproduce this on 5.2-rel or -current. Can you > > provide the options used to iozone? > > I think I used the defaults other than cranking up the max file size go > 1GB to match the memory size on my server. It took a long time to run > the full set of tests ... OK, I'll tive it a try. NFS between different FreeBSD versions in my pile-o-machinery have all come up OK, but I'll try against redhat. -- Doug White | FreeBSD: The Power to Serve dwhite_at_gumbysoft.com | www.FreeBSD.orgReceived on Tue Jan 13 2004 - 08:03:50 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:37 UTC