It seems as though we've been having problems with our filesystem server
ever since we went to XFS. Everytime we try doing anything large on it,
the kupdated seems to take a heavy load, and nfsd seems to freeze up on
us. Any ideas? We are using 2.4.18-4SGI_XFS_1.1smp.
ps xwo pid,command,wchan gives us the following (there is more but I
cropped some of the nfsd schedule_timeouts out).
PID COMMAND WCHAN
1 init [3] schedule_timeout
2 [keventd] context_thread
3 [ksoftirqd_CPU0] ksoftirqd
4 [ksoftirqd_CPU1] ksoftirqd
5 [kswapd] schedule_timeout
6 [bdflush] bdflush
7 [kupdated] schedule_timeout
8 [pagebuf_daemon] pagebuf_daemon
9 [mdrecoveryd] md_thread
15 [scsi_eh_0] down_interruptible
16 [scsi_eh_1] down_interruptible
18 [scsi_eh_2] down_interruptible
652 syslogd -m 0 schedule_timeout
657 klogd -2 do_syslog
857 /usr/sbin/sshd schedule_timeout
890 xinetd -stayaliv schedule_timeout
900 /opt/arkeia/bin/ schedule_timeout
941 rpc.rquotad schedule_timeout
946 rpc.mountd schedule_timeout
951 [nfsd] -
952 [nfsd] schedule_timeout
953 [nfsd] schedule_timeout
990 [nfsd] down
991 [nfsd] schedule_timeout
992 [nfsd] schedule_timeout
993 [lockd] schedule_timeout
994 [rpciod] end
1012 /usr/sbin/dhcpd schedule_timeout
1048 crond schedule_timeout
1159 /sbin/mingetty t schedule_timeout
1160 /sbin/mingetty t schedule_timeout
1161 /sbin/mingetty t schedule_timeout
1162 /sbin/mingetty t schedule_timeout
1163 /sbin/mingetty t schedule_timeout
1164 /sbin/mingetty t schedule_timeout
1167 /usr/sbin/sshd schedule_timeout
1169 -bash wait4
1246 ps xwo pid,comma -
I'd appreciate any input.
Thank you for your time,
Mako Furukawa
|