Hello, I'm having a hard time trying to identify what produces the iowait to go up to 90-95% (leaving 0% idle) for a while (a couple of seconds to a few minutes) and then descend to 5% max. (I've already readed the similar subject thread)

This is a linux:

Linux 2.4.21-32.ELsmp #1 SMP Fri Apr 15 21:17:59 EDT 2005 i686 i686 i386 GNU/Linux

where runs an Oracle RDBMS 10gr2 server

with this cpu's:

[root@abcdb root]# cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 15
model : 4
model name : Intel® Xeon™ CPU 3.20GHz
stepping : 1
cpu MHz : 3200.272
cache size : 1024 KB
physical id : 0
siblings : 2
runqueue : 0
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 5
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm lm
bogomips : 6383.20

processor : 1
vendor_id : GenuineIntel
cpu family : 15
model : 4
model name : Intel® Xeon™ CPU 3.20GHz
stepping : 1
cpu MHz : 3200.272
cache size : 1024 KB
physical id : 0
siblings : 2
runqueue : 0
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 5
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm lm
bogomips : 6396.31

and this hd in /:

hdparm -tT /dev/cciss/c0d0p3

/dev/cciss/c0d0p3:
Timing buffer-cache reads: 3196 MB in 2.00 seconds = 1598.00 MB/sec
Timing buffered disk reads: 216 MB in 3.02 seconds = 71.52 MB/sec

and this io:

[root@abcdb root]# iostat 1
Linux 2.4.21-32.ELsmp 09/29/2006

avg-cpu: %user %nice %sys %iowait %idle
14.02 0.13 1.48 8.33 76.03

Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
cciss/c0d0 29.65 44.41 381.73 290819044 2499835932
cciss/c0d0p1 0.00 0.00 0.00 3170 1332
cciss/c0d0p2 4.60 30.51 34.96 199824280 228948688
cciss/c0d0p3 25.05 13.89 346.77 90991330 2270885912
cciss/c0d1 19.48 450.39 23.00 2949492098 150653464
cciss/c0d1p1 19.48 450.39 23.00 2949491842 150653464

avg-cpu: %user %nice %sys %iowait %idle
1.01 0.00 0.00 9.05 89.95

Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
cciss/c0d0 14.00 112.00 0.00 112 0
cciss/c0d0p1 0.00 0.00 0.00 0 0
cciss/c0d0p2 0.00 0.00 0.00 0 0
cciss/c0d0p3 14.00 112.00 0.00 112 0
cciss/c0d1 1.00 0.00 8.00 0 8
cciss/c0d1p1 1.00 0.00 8.00 0 8

avg-cpu: %user %nice %sys %iowait %idle
0.00 0.00 0.00 4.00 96.00

Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
cciss/c0d0 0.00 0.00 0.00 0 0
cciss/c0d0p1 0.00 0.00 0.00 0 0
cciss/c0d0p2 0.00 0.00 0.00 0 0
cciss/c0d0p3 0.00 0.00 0.00 0 0
cciss/c0d1 5.00 16.00 96.00 16 96
cciss/c0d1p1 5.00 16.00 96.00 16 96

do you have any idea? (thank you!)

It depends on what's going on at that specific time. Try recording output of "ps wauxf" at a time when there's a high IO_WAIT and try to look for suspicious processes that take up much CPU.

Upgrading your kernel to a 2.6.15+ version should be helpful a bit, but not a very big deal of difference.