Iostat Result Analysis
[Kefu@sz-8 linux]$ iostat-x-K
Linux 2.6.18-128.el5_cyou_1.0 (SZ-8.30) 09/08/2011
AVG-CPU:%user%nice%system%iowait%steal%idle
16.58 0.00 2.79 0.46 0.00 80.16
device:rrqm/s wrqm/s r/s w/s rkb/s wkb/s avgrq-sz avgqu-sz await SVCTM%util
SDA 0.06 29.28 0.22 37.14 10.21 265.68 14.77 0.02 0.51 0.15 0.55
SDA1 0.00 0.00 0.00 0.00 0.00 0.00 10.79 0.00 2.66 2.43 0.00
Sda2 0.01 0.78 0.10 0.36 0.81 4.58 23.51 0.00 1.21 0.84 0.04
Sda3 0.03 15.17 0.09 35.39 8.98 202.24 11.91 0.01 0.26 0.12 0.44
SDA4 0.00 0.00 0.00 0.00 0.00 0.00 2.00 0.00 33.33 33.33 0.00
Sda5 0.01 1.59 0.03 0.51 0.34 8.40 32.20 0.00 1.19 0.58 0.03
Sda6 0.00 0.00 0.00 0.12 0.00 0.48 8.18 0.00 5.02 4.53 0.05
SDA7 0.00 0.00 0.00 0.00 0.00 0.00 45.00 0.00 5.52 3.04 0.00
SDA8 0.00 0.00 0.00 0.00 0.00 0.00 40.88 0.00 7.62 6.03 0.00
SDA9 0.00 0.00 0.00 0.00 0.00 0.00 39.71 0.00 7.37 5.83 0.00
SDA10 0.00 0.00 0.00 0.00 0.00 0.00 37.57 0.00 5.70 3.54 0.00
SDA11 0.00 11.74 0.01 0.76 0.08 49.97 131.48 0.01 10.74 0.57 0.04
SDB 0.01 3.91 20.24 20.21 1262.95 1853.94 154.09 0.52 12.84 1.97 7.95
RRQM/S: Number of read operations per second for merge. Delta (rmerge)/s
wrqm/s: Number of write operations per second for merge. Delta (wmerge)/s
R/S: Number of Read I/O devices completed per second. Delta (RIO)/s
W/S: Number of write i/0 devices completed per second. Delta (WIO)/s
RSEC/S: Number of sectors read per second. Delta (rsect)/s
WSEC/S: Number of sector writes per second. Delta (wsect)/s
RKB/S: The number of K bytes read per second. Is half the rsec/s, because each sector size is 512 bytes.
WKB/S: The number of K bytes written per second. It's half the wsec/s.
Avgrq-sz: The average data size (sector) per device I/O operation. Delta (rsect+wsect)/delta (Rio+wio)
Avgqu-sz: Average I/O queue length. That is, Delta (AVEQ)/s/1000 (because the Aveq unit is in milliseconds)
Await: The average wait time (in milliseconds) for each device I/O operation. Delta (ruse+wuse)/delta (Rio+wio)
SVCTM: Average service time (in milliseconds) per device I/O operation. Delta (use)/delta (RIO+WIO)
%util: How much time is spent in a second for I/O operations, or how many times in a second I/O queues are non-empty. Delta (USR)/s/1000 (because the unit of use is in milliseconds)
If%util is close to 100%, there are too many I/O requests, the I/O system is full, and the disk may have a bottleneck.
The more important parameters
%util: How much time is spent in a second for I/O operations, or how many times in a second I/O queues are non-null
SVCTM: Average service time for each device I/O operation
Await: Average wait time for each device I/O operation
Avgqu-sz: Average I/O Queue Length
If%util close to 100%, indicating I/O request too much, I/O system is full load, disk may have bottlenecks, general%util greater than 70%,i/o pressure is relatively large, reading speed more than wait.
You can also combine vmstat to view the B parameter (the number of processes waiting for the resource) and the WA parameter (percentage of CPU time taken by I/O wait, higher I/O pressure over 30%)
The size of the await generally depends on the service time (SVCTM) and the length of the I/O queue and the emit mode of I/O requests. If the SVCTM is closer to await, there is almost no wait time for I/O;
Await is much larger than SVCTM, indicating that the I/O queues are too long and that the response time is slower to apply.
The metaphor of the image
r/s+w/s similar to the total number of people who paid
Average Queue Length (AVGQU-SZ) similar to the average number of people queuing in a unit of time
Average service time (AVCTM) similar to Cashier's collection speed
Average wait time (await) is similar to the average waiting time per person
Average I/O data (AVGRQ-SZ) is similar to what everyone buys on average
I/O operating rate (%util) is similar to the percentage of time someone queues in a cashier
SVCTM are typically less than await (because waiting times for waiting requests are repeatedly computed), SVCTM size is generally related to disk performance, and cpu/memory load can also affect it, too many requests
Indirectly lead to the increase of SVCTM. The size of the await generally depends on the service time (SVCTM) and the length of the I/O queue and the emit mode of I/O requests. If the SVCTM is closer to await, it means that I/O has little
Wait time; If the await is much larger than the SVCTM, the I/O queue is too long, the response time is slow, and if the response time exceeds the allowable scope of the user, then you can consider replacing the faster disk, adjust
Whole kernel elevator algorithm, optimize application, or upgrade CPU
Queue Length (AVCQU-SZ) can also be used as an indicator of system I/O load, but since Avcqu-sz is average per unit time, it does not reflect instantaneous I/O floods.