1

I have a server where iostat shows weird numbers to me:

avg-cpu:  %user   %nice %system %iowait  %steal   %idle
          22.5%    0.0%    0.8%    1.3%    0.0%   75.5%

Device             tps    kB_read/s    kB_wrtn/s    kB_read    kB_wrtn
loop0
                  0.00         0.0k         0.0k       0.0k       0.0k
loop1
                  0.00         0.0k         0.0k       0.0k       0.0k
sda
                  1.40         6.4k        64.2k      32.0k     321.0k
sdb
                549.40        11.8M        83.9M      59.2M     419.6M
dm-0
                680.80        11.8M       220.5M      58.9M       1.1G

sdb is under dm-0 device, but how it is possible that dm-0 has more data written on it that sdb. In my opinion, they must always match and this is true for reading operations. I'm confused and don't know which numbers to believe.

Dmitry Maslennikov
  • 66
  • 2

1 Answers1

0

Probably it's page caching of your operating system. Apply a write load on dm-0 while running this command in an other terminal window:

watch -n.5 "cat /proc/meminfo |grep -i dirty | awk {'print$2$3'}"

If it grows, then your write requests are witten in memory for later eviction to actual physical disk. For more info on page cache: Page Cache

Far Had
  • 11
  • 2