TC358. Actual speed is less than expected

[VenyaPanou]

Actual speed : root@vm:~# fio fiorandomreadlatency.fio job1: (g=0): rw=randread, bs=(R) 4096B-4096B, (W) 4096B-4096B, (T) 4096B-4096B, ioengine=libaio, iodepth=1 fio-3.16 Starting 1 process Jobs: 1 (f=1): [r(1)][100.0%][r=10.7MiB/s][r=2742 IOPS][eta 00m:00s] job1: (groupid=0, jobs=1): err= 0: pid=1227: Mon Jun 27 06:53:37 2022 read: IOPS=2732, BW=10.7MiB/s (11.2MB/s)(1281MiB/120001msec) slat (usec): min=11, max=600, avg=29.40, stdev=10.56 clat (usec): min=5, max=5490, avg=329.38, stdev=61.16 lat (usec): min=204, max=5565, avg=359.67, stdev=61.84 clat percentiles (usec): | 1.00th=[ 249], 5.00th=[ 269], 10.00th=[ 277], 20.00th=[ 285], | 30.00th=[ 293], 40.00th=[ 306], 50.00th=[ 314], 60.00th=[ 326], | 70.00th=[ 343], 80.00th=[ 371], 90.00th=[ 408], 95.00th=[ 441], | 99.00th=[ 494], 99.50th=[ 523], 99.90th=[ 619], 99.95th=[ 750], | 99.99th=[ 1598] bw ( KiB/s): min= 8432, max=12288, per=99.98%, avg=10926.26, stdev=628.41, samples=239 iops : min= 2108, max= 3072, avg=2731.53, stdev=157.12, samples=239 lat (usec) : 10=0.01%, 20=0.01%, 50=0.01%, 100=0.01%, 250=1.09% lat (usec) : 500=98.07%, 750=0.79%, 1000=0.02% lat (msec) : 2=0.03%, 4=0.01%, 10=0.01% cpu : usr=3.68%, sys=9.07%, ctx=335723, majf=0, minf=11 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=327870,0,0,0 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): READ: bw=10.7MiB/s (11.2MB/s), 10.7MiB/s-10.7MiB/s (11.2MB/s-11.2MB/s), io=1281MiB (1343MB), run=120001-120001msec

[xmonader]

please attach the fio script

[xmonader]

could be related to https://github.com/threefoldtech/zos/issues/1467 @muhamadazmy

[dmitri-alleo]

@xmonader

The configuration

fiorandomreadlatency.fio (file name)
[global]
bs=4K
iodepth=1
direct=1
ioengine=libaio
group_reporting
time_based
runtime=120
numjobs=1
name=readlatency-test-job
rw=randread
filesize=10K

[job1]
filename=device name

And then we use it via

fio fiorandomreadlatency.fio

[muhamadazmy]

Can you please provide information about your workloads:

• Flist used
• Node id, and network id.

It's known that when using containers flist performance of the rootfs of the container is degraded. Instead data disk needed to be added to the container (rootfs is only to serve your image files). or instead use a VM image.