1

I'm trying to debug some weird behaviour with Postgres performance and am trying to understand the iostat output. I'm on 2x SSD on RAID1. Why would one of the disks be reported at ~21% utilisation and the other at ~95% utilisation? Shouldn't they be exactly the same?

Could this in any way be related to this question?

avg-cpu:  %user   %nice %system %iowait  %steal   %idle
          10.03    0.00    5.28    1.24    0.00   83.45

Device            r/s     w/s     rMB/s     wMB/s   rrqm/s   wrqm/s  %rrqm  %wrqm r_await w_await aqu-sz rareq-sz wareq-sz  svctm  %util
loop0            0.00    0.00      0.00      0.00     0.00     0.00   0.00   0.00    0.00    0.00   0.00     0.00     0.00   0.00   0.00
sda             83.20  493.60      0.72      8.11     0.00     1.80   0.00   0.36    0.56    0.39   0.24     8.83    16.83   0.37  21.36
sdb             24.40  493.60      0.24      8.11     0.00     1.80   0.00   0.36    3.41    1.91   1.02    10.10    16.83   1.82  94.24
md0              0.00    0.00      0.00      0.00     0.00     0.00   0.00   0.00    0.00    0.00   0.00     0.00     0.00   0.00   0.00
md1              0.00    0.00      0.00      0.00     0.00     0.00   0.00   0.00    0.00    0.00   0.00     0.00     0.00   0.00   0.00
md2            107.60  494.60      0.96      8.12     0.00     0.00   0.00   0.00    0.00    0.00   0.00     9.12    16.81   0.00   0.00
Saurabh Nanda
  • 449
  • 1
  • 7
  • 17
  • With regards to `%util` the manual [`man iostat`](http://man7.org/linux/man-pages/man1/iostat.1.html) gives a big caveat with regards to SSD's: *"for devices serving requests in parallel, such as RAID arrays and modern SSDs, this number does not reflect their performance limits.`"* - so by itself it is hard to say if those different numbers are significant or not. Maybe take a look at the SMART health status of the drives? – HBruijn Jan 28 '19 at 11:16
  • @HBruijn I'm just surprised why both the drives would be reporting such a big difference in %util. I'm not necessarily surprised by the absolute values - because of the caveat that you mentioned. – Saurabh Nanda Jan 28 '19 at 11:27
  • @HBruijn what do I check in the SMART health status exactly, and how? – Saurabh Nanda Jan 28 '19 at 11:29
  • [`smartctl -xall /dev/sda`](https://linux.die.net/man/8/smartctl) resp. /dev/sdb – HBruijn Jan 28 '19 at 12:01

0 Answers0