Skip to contentSkip to conversation

system time/2.4.18-3smp/3ware

DISCUSSION Larry McVoy 1 participant 1 message Feb 22, 2003

lore
Share & tools
lore
#0 lore raw

This is watching 3 drives being tar-ed | wc -c in parallel (I suspect a bad drive, trying to narrow it down). The machine is a dual 1.6ghz Athlon, Tyan motherboard, 2GB of ram.

I'm a little surprised at the system time - yeah, 72MB/sec is a far amount of I/O but should we really be eating up that much system time?

load free cach swap pgin  pgou dk0 dk1 dk2 dk3 ipkt opkt  int  ctx  usr sys idl
3.16  11M 1.4G   0   74M  2.4K   0   0   0   0    0    0  1.3K  46K   0  92 108
3.23  11M 1.4G   0   71M  2.4K   0   0   0   0    0    0  1.3K  45K   1  88 111
3.29  11M 1.4G   0   72M  2.4K   0   0   0   0    0    0  1.2K  46K   2 105  93
3.35  11M 1.4G   0   73M  3.2K   0   0   0   0    0    0  1.2K  45K   1 100  99
3.32  11M 1.4G   0   74M  2.4K   0   0   0   0    0    0  1.2K  45K   1  98 101
3.53  11M 1.4G   0   74M   65K   0   0   0   0    0    0  1.3K  46K   4 103  93

I thought maybe it was the | wc -c (notice the context switches/sec) so I tried it with a cpio into /dev/null (tar neatly "optimizes" out any I/O when you are going to /dev/null, I really hate that) and it was still high

load free cach swap pgin  pgou dk0 dk1 dk2 dk3 ipkt opkt  int  ctx  usr sys idl
0.78  12M 1.4G   0   66M   34K   0   0   0   0    0    0  945  2.0K   6  82 112
0.96  11M 1.4G   0   66M  2.4K   0   0   0   0  1.0  1.0  964  2.1K   6  79 115
1.12  11M 1.4G   0   66M  2.4K   0   0   0   0    0    0  1.0K 2.3K   7  82 111
1.27  11M 1.4G   0   66M  2.4K   0   0   0   0    0    0  972  2.1K   4  80 116
1.41  11M 1.4G   0   67M  2.4K   0   0   0   0    0    0  960  2.1K   5  87 108
1.62  11M 1.4G   0   67M  2.4K   0   0   0   0    0    0  1.0K 2.3K   7  83 110
Signature · 2 lines
---
Larry McVoy            	 lm at bitmover.com           http://www.bitmover.com/lm