Well, this machine is 11 years old now. This explains the large amount of CPU time.
Quad 900MHz UltraSparc III processors are more than enough to handle a simple filesystem.
The server runs fine, is patched up to date. The UFS filesystem that
was
used is actually the root filesystem and it is a metadevice mirror of the two internal disks.
A simple mirror is slow.
I don't agree.
jupiter-sparc-SunOS5.10 # pwd /usr/local/src jupiter-sparc-SunOS5.10 # which mkfile /usr/sbin/mkfile
jupiter-sparc-SunOS5.10 # ptime /usr/sbin/mkfile 1024m one_gig.dat
real 30.874 user 0.206 sys 5.574 jupiter-sparc-SunOS5.10 # ls -l total 3035664 -rw-r--r-- 1 root root 479733760 Aug 9 15:44 linux-3.5.1.tar -rw------T 1 root root 1073741824 Aug 13 20:00 one_gig.dat jupiter-sparc-SunOS5.10 # rm one_gig.dat jupiter-sparc-SunOS5.10 #
Meanwhile in another xterm I ran iostat :
$ iostat -xc -d ssd0 ssd2 5 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.2 4.8 1.5 46.2 0.0 0.1 34.2 0 3 1 2 0 97 ssd2 0.1 4.6 1.2 46.1 0.0 0.1 34.7 0 3 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.0 9.2 0.0 5.0 0.0 0.2 20.7 0 5 0 1 0 99 ssd2 0.0 6.4 0.0 3.6 0.0 0.1 15.5 0 3 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 0 1 0 99 ssd2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.4 111.4 2.0 12546.4 6.3 4.3 94.9 25 36 1 2 0 97 ssd2 0.0 104.6 0.0 12476.0 5.9 4.1 96.1 24 31 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.8 326.4 6.4 37426.6 28.6 13.1 127.4 80 100 0 6 0 93 ssd2 0.0 317.8 0.0 37423.8 26.4 12.5 122.6 76 91 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.2 322.6 1.6 35021.9 23.2 12.5 110.7 73 98 0 17 0 83 ssd2 1.2 312.8 9.6 35104.7 23.0 12.5 113.1 73 92 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 2.2 294.8 20.8 30385.4 16.8 11.9 96.8 64 98 0 11 0 89 ssd2 1.4 284.8 11.2 30418.4 16.1 11.2 95.4 63 90 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.0 314.0 0.0 33055.0 20.3 12.2 103.6 69 98 0 5 0 95 ssd2 1.8 300.2 14.4 32987.6 19.1 11.7 102.0 67 89 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.4 312.8 3.2 33398.0 26.7 12.8 126.0 75 98 0 6 0 94 ssd2 1.6 304.4 12.8 33424.0 24.0 11.9 117.1 71 89 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 2.4 278.6 19.2 28227.5 24.8 10.8 126.9 63 94 1 7 0 92 ssd2 1.0 259.2 8.0 28187.7 21.3 9.6 118.6 57 75 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.0 9.4 0.0 27.6 0.0 0.2 25.1 0 6 0 1 0 99 ssd2 0.0 6.6 0.0 26.2 0.0 0.1 15.3 0 3 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.0 18.0 0.0 62.3 0.0 0.4 21.2 0 11 0 1 0 98 ssd2 0.2 13.2 0.2 59.9 0.0 0.2 16.8 0 7 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.4 29.0 3.2 307.1 0.0 0.8 26.2 0 15 0 2 0 98 ssd2 0.2 29.0 1.6 306.7 0.0 0.8 29.1 0 15 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.0 0.8 0.0 0.4 0.0 0.0 10.0 0 1 0 5 0 95 ssd2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0 extended device statistics cpu device r/s w/s kr/s kw/s wait actv svc_t %w %b us sy wt id ssd0 0.0 1.8 0.0 1.2 0.0 0.0 7.8 0 1 0 1 0 99 ssd2 0.0 0.6 0.0 0.6 0.0 0.0 7.1 0 0 ^C$ $
bulk throughput would be 1024m/30.874sec =
jupiter-sparc-SunOS5.10 # dc 9k 1024 30.874 / p 33.167066139
Not bad.
However, a 2300 GB FCAL drive should be faster.
Probably, but not much.
But... did you turn on logging?
ufs logging ? Yes. Always. However that would be the same with the gtar and star tests also.
Dennis