How to Measure IOPS Using fio

It is important that your Swimlane instance meets the minimum required IOPS, otherwise severe performance degradation can result. Also, you risk your instance being unsupported by Swimlane. This article describes one way to measure IOPS using the fio tool. Note: Because of file caching, the results of fio can vary depending on when you run the command. To reduce that variance, we recommend running fio several times a day over a couple of days and averaging the result. Install fio if not already installed: Fedora-based Linux (e.g. Red Hat, CentOS)

$ sudo yum install fio

Debian-based Linux (e.g. Ubuntu)

$ sudo apt-get install fio

Then run this command on each of your cluster nodes. It will run for a few minutes with output that updates in real time; when it's finished it, will have final output similar to this:

$ fio --name=random-write --ioengine=posixaio --rw=randwrite --bs=4k --numjobs=1 --size=4g --iodepth=1 --runtime=60 --time_based --end_fsync=1 random-write: (g=0): rw=randwrite, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=posixaio, iodepth=1 fio-3.7 Starting 1 process random-write: Laying out IO file (1 file / 4096MiB) Jobs: 1 (f=1): [F(1)][100.0%][r=0KiB/s,w=0KiB/s][r=0,w=0 IOPS][eta 00m:00s] random-write: (groupid=0, jobs=1): err= 0: pid=4168: Thu Aug 19 11:09:57 2021 write: IOPS=3994, BW=15.6MiB/s (16.4MB/s)(2449MiB/156935msec) slat (nsec): min=1681, max=105529k, avg=4310.96, stdev=142005.21 clat (nsec): min=820, max=228958k, avg=89450.02, stdev=1359359.52 lat (usec): min=11, max=228961, avg=93.76, stdev=1370.29 clat percentiles (usec): | 1.00th=[ 11], 5.00th=[ 12], 10.00th=[ 14], 20.00th=[ 16], | 30.00th=[ 19], 40.00th=[ 20], 50.00th=[ 20], 60.00th=[ 21], | 70.00th=[ 23], 80.00th=[ 29], 90.00th=[ 36], 95.00th=[ 48], | 99.00th=[ 130], 99.50th=[ 791], 99.90th=[21103], 99.95th=[28967], | 99.99th=[50070] bw ( KiB/s): min= 3787, max=227400, per=100.00%, avg=41644.82, stdev=54714.18, samples=120 iops : min= 946, max=56850, avg=10411.00, stdev=13678.61, samples=120 lat (nsec) : 1000=0.01% lat (usec) : 2=0.01%, 4=0.01%, 10=0.01%, 20=52.57%, 50=42.84% lat (usec) : 100=3.25%, 250=0.65%, 500=0.13%, 750=0.05%, 1000=0.04% lat (msec) : 2=0.08%, 4=0.05%, 10=0.09%, 20=0.13%, 50=0.10% lat (msec) : 100=0.01%, 250=0.01% cpu : usr=2.06%, sys=2.00%, ctx=627931, majf=0, minf=50 IO depths : 1=100.0%, 2=0.0%, 4=0.0%, 8=0.0%, 16=0.0%, 32=0.0%, >=64=0.0% submit : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% complete : 0=0.0%, 4=100.0%, 8=0.0%, 16=0.0%, 32=0.0%, 64=0.0%, >=64=0.0% issued rwts: total=0,626892,0,1 short=0,0,0,0 dropped=0,0,0,0 latency : target=0, window=0, percentile=100.00%, depth=1 Run status group 0 (all jobs): WRITE: bw=15.6MiB/s (16.4MB/s), 15.6MiB/s-15.6MiB/s (16.4MB/s-16.4MB/s), io=2449MiB (2568MB), run=156935-156935msec Disk stats (read/write): dm-0: ios=453/462780, merge=0/0, ticks=1795/7586994, in_queue=7615234, util=93.05%, aggrios=453/460952, aggrmerge=0/1823, aggrticks=1795/7571606, aggrin_queue=7572764, aggrutil=92.87% sda: ios=453/460952, merge=0/1823, ticks=1795/7571606, in_queue=7572764, util=92.87%

Using this piece of the output: ”… WRITE: bw=15.6MiB/s (16.4MB/s), 15.6MiB/s-15.6MiB/s (16.4MB/s-16.4MB/s)… …” convert the number in bold to IOPS using this formula: (MBps Throughput / 4) * 1024. In this example: IOPS = (16.4/4) * 1024 = 4198.4. If the bold number is less than 11.7 MB/s, then you do not meet the minimum required 3000 IOPS.