In this post are the results of running a few more benchmarking tests on the KVM VPS’es machines. In particular, running the dd test and also a failed attempt at running bench.sh.

Results of ‘dd’ test

Command: sudo dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

Stockservers, Frankfurt, Germany

16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 24.6735 s, 43.5 MB/s

This is the result that surprised me the most, so I ran another test, with similar result. An internet search led me to this article on cyberciti dot biz, so I ran two additional runs, as follows:

Running hdparm on SSD on a KVM VPS running Debian 9. Screenshot by Amar Vyas
Running hdparm on SSD – VPS

Run # 3
[email protected]:/tmp$ dd if=/dev/zero of=/tmp/test.bin bs=1G count=1 oflag=direct
1+0 records in
1+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.55136 s, 692 MB/s

Running dd test to check disk's speed
Running dd test to check disk’s speed

Run # 4
[email protected]:/tmp$ dd if=/dev/zero of=/tmp/test2.img bs=512 count=1000 oflag=dsync
1000+0 records in
1000+0 records out
512000 bytes (512 kB, 500 KiB) copied, 0.879881 s, 582 kB/s

Hmm… why is this happening? now that makes my homework for the weekend 🙂

Added later: I also read up on the results of the dd test Lifewire, as well as here and here . Something is definitely odd with the results on Stockservers VPS. Time to raise a ticket!
Reinstalling the Operating System on the Stockservers VPs seemed to have helped: here are the results :
Run # 5

sudo dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync

16384+0 records in

16384+0 records out

1073741824 bytes (1.1 GB, 1.0 GiB) copied, 4.00194 s, 268 MB/s


Hostslick, The Netherlands

16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 5.27266 s, 204 MB/s


AWS,Mumbai, India

16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 7.93384 s, 135 MB/s


Ramnode, Dallas, USA

16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.79779 s, 597 MB/s


Hostdoc, Finland

16384+0 records in

16384+0 records out

1073741824 bytes (1.1 GB, 1.0 GiB) copied, 34.4042 s, 31.2 MB/s


Hosthatch, Chicago

16384+0 records in

16384+0 records out

1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.49644 s, 718 MB/s


Bench script: bench.sh

Surprisingly, bench.sh script resulted in hanging once the speed tests began to run. Here are the results:

Source: https://github.com/haydenjames/bench-scripts
Command:
wget https://raw.githubusercontent.com/hidden-refuge/bench-sh-2/master/bench.sh && chmod +x bench.sh && ./bench.sh
AWS
Location Provider Speed
CDN Cachefly 31.2MB/s
Atlanta, GA, US Coloat 3.99MB/s
Dallas, TX, US Softlayer 4.51MB/s
Seattle, WA, US Softlayer 4.97MB/s
San Jose, CA, US Softlayer 4.45MB/s
—-
Hostslick
Location Provider Speed
CDN Cachefly 63.2MB/s
Atlanta, GA, US Coloat 3.86MB/s
Dallas, TX, US Softlayer 8.20MB/s
Seattle, WA, US Softlayer 6.87MB/s
San Jose, CA, US Softlayer 5.70MB/s
Ramnode
Location Provider Speed
CDN Cachefly 55.3MB/s
Atlanta, GA, US Coloat 29.4MB/s
Dallas, TX, US Softlayer 56.2MB/s
Seattle, WA, US Softlayer 12.5MB/s
San Jose, CA, US Softlayer 14.7MB/s
Hosthatch
Location Provider Speed
CDN Cachefly 87.1MB/s
Atlanta, GA, US Coloat 616KB/s
Dallas, TX, US Softlayer 55.2MB/s
Seattle, WA, US Softlayer 36.9MB/s
San Jose, CA, US Softlayer 26.5MB/s
Stockservers
Location Provider Speed
CDN Cachefly 56.6MB/s
Atlanta, GA, US Coloat 5.84MB/s
Dallas, TX, US Softlayer 7.93MB/s
Seattle, WA, US Softlayer 9.85MB/s
San Jose, CA, US Softlayer 9.21MB/s
—-
To sum it up: Learnt a thing or two in the process, and the learning continues. I will look into the Stockservers Disk performance. Particularly considering I need to decide in a week’s time whether to renew or not.
Running-Benchmark Tests on VPS - screenshot by Amar Vyas, Sept 2019
Running-Benchmark Tests on VPS

Leave a Reply

Your email address will not be published. Required fields are marked *

Name *