Question & Answer
Question
How do I generate IOSTAT output on Linux and determine if there are disk performance problems?
Cause
Disk performance
Answer
On Linux you can issue the command:
/usr/bin/iostat -xtk 30 3
This will gather 3 collections at 30 second intervals. Change the interval or number of collections as needed or requested.
Sample output will look similar to the following:
08/04/2015 11:07:49 AM
avg-cpu: %user %nice %system %iowait %steal %idle
1.42 0.00 4.14 2.26 0.00 92.17
Device: rrqm/s wrqm/s r/s w/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util
sda 0.01 2.78 0.04 1.55 0.96 17.31 23.08 0.00 0.86 0.24 0.04
sdb 0.00 0.00 0.00 0.07 0.06 0.64 20.52 0.00 2.91 2.80 0.02
dm-40 0.00 0.00 16.79 14.42 4297.73 3645.84 509.11 0.06 2.06 1.88 5.88
dm-41 0.00 0.00 16.17 14.29 4138.88 3614.98 509.09 0.06 2.00 1.84 5.60
dm-42 0.00 0.00 16.73 14.47 4281.44 3656.40 508.91 0.06 2.00 1.85 5.76
dm-43 0.00 0.00 16.04 13.71 4105.63 3466.60 509.11 0.06 1.99 1.84 5.48
dm-44 0.00 0.00 19.16 19.08 4890.17 4784.78 506.02 0.11 2.95 2.72 10.40
dm-45 0.00 0.00 18.20 18.17 4643.35 4553.89 505.76 0.09 2.60 2.41 8.77
dm-46 0.00 0.00 18.53 18.89 4727.78 4733.02 505.65 0.12 3.08 2.82 10.54
dm-47 0.00 0.00 18.27 18.49 4661.67 4632.69 505.75 0.10 2.59 2.41 8.85
dm-48 0.00 0.00 17.79 18.07 4540.63 4525.51 505.59 0.07 2.04 1.94 6.97
dm-49 0.00 0.00 17.63 17.95 4498.44 4498.54 505.71 0.07 2.04 1.95 6.95
The primary statistics to review are the await and %util. We look for the "await" to be 5ms or less and the %util to be 70% or less.
Occasional spiking may occur however sustained values above 5ms and/or 70% utilization can indicate a disk problem.
Product Synonym
TSM ADSM ITSM Tivoli Storage Manager IBM Spectrum Protect
Was this topic helpful?
Document Information
Modified date:
17 June 2018
UID
swg21978034