[linux]-linux通过diskstats来计算磁盘写入的延迟

导语：看到资料有计算磁盘延迟的。写一个类似的shell脚本 以防后续要用。

主要通过命令 cat /proc/diskstats 来获取磁盘的信息

    fields:
     1 - major number
     2 - minor mumber
     3 - device name
     4 - reads completed successfully
     5 - reads merged
     6 - sectors read
     7 - time spent reading (ms)
     8 - writes completed
     9 - writes merged
    10 - sectors written
    11 - time spent writing (ms)
    12 - I/Os currently in progress
    13 - time spent doing I/Os (ms)
    14 - weighted time spent doing I/Os (ms)

# 第七列为 读花的时间（ms），这是所有读操作所花费的毫秒数（用__make_request()到end_that_request_last()测量）

# 第十三列为 输入输入花的时间（ms），花在I/O操作上的毫秒数，这个域会增长只要field 9不为0。
（number of milliseconds spent doing I/Os. This field is increased so long as field 9 is nonzero.）

cat /proc/diskstats 

   8       0 sda 208010923 6190 51045399672 78018748 40801137 2919609 20073260568 1403219000 0 49226220 1481074792
   8       1 sda1 208010836 6190 51045397248 78018568 40801137 2919609 20073260568 1403219000 0 49225668 1481044656
   8      16 sdb 35632370 1766 2055298292 83563540 11908017 4001307 3088071384 791169568 0 20208620 874718584

通过获取磁盘60秒内写盘花的时间(毫秒) 和写磁盘的次数 相除来获得每次写需要的耗时。

1分钟内写操作花费的毫秒数/写入完成 iops的数量

while :
do
#    clear
    current_time=`date "+%Y-%m-%dT%H:%M:%S"` #当前时间
    echo current_time: $current_time
    current_time_spent_reading=`cat /proc/diskstats |grep sdb |awk   '{print $11}' |tail -1`
    echo current_time_spent_reading $current_time_spent_reading
    current_writes_completed=`cat /proc/diskstats |grep sdb |awk   '{print $8}'|tail -1`
    echo current_writes_completed $current_writes_completed
    sleep 60

    time_now=`date "+%Y-%m-%dT%H:%M:%S"` #当前时间
    echo time_now $time_now
    time_spent_reading_now=`cat /proc/diskstats |grep sdb |awk   '{print $11}'|tail -1`
    echo time_spent_reading_now $time_spent_reading_now
    writes_completed_now=`cat /proc/diskstats |grep sdb |awk   '{print $8}'|tail -1`
    echo writes_completed_now $writes_completed_now

    if [ "$current_writes_completed" != "$writes_completed_now" ];then
      write_counts=$(($writes_completed_now - $current_writes_completed))
      #aaa=$(($a - $b))
      time_spent_reading=$(($time_spent_reading_now - $current_time_spent_reading))
      #$[num1 - num2]    $[num1 / num2]
      delay_time=$(($time_spent_reading / $write_counts))
      echo ------------
      echo write_counts $write_counts
      echo time_spent_reading $time_spent_reading
      echo  delay_time $delay_time ms
      echo ++++++++++++
   else
      echo "没有写"
   fi
done

测试

cat /proc/diskstats | grep sdb1

dd if=kube-bench of=/dev/null

# 测试方式：使用dd指令，对磁盘进行连续写入，不使用内存缓冲区，每次写入8k的数据，总共写入20万次，产生1.6G大小的文件。
# if=FILE         read from FILE instead of stdin   of=FILE         write to FILE instead of stdout
# 测试指令：dd if=/dev/zero of=/data01/test.dbf bs=8k count=200000 conv=fdatasync


cat /proc/diskstats | grep sdb1