Contabo Review - VPS M SSD

Ympker

Updated Review

I feel like I have to row back a bit and give kudos to Contabo for what its worth. After talking to a Contabo Rep on the phone regarding my issues and concerns the problem which was most annoying to me (the limited SSD speed) has been fixed and is working quite good right now.
That being said I need to say it is good to know that Contabo is willing to talk things through and act when needed. Coincidence or triggered by my reviews? Even if the latter applies, from a Salesman POV customers who do complain and acctually talk about their problems are always getting treated accordingly. This does not only apply to Contabo but probably any company out there.

This experience is about Contabo.de (although I doubt contabo.com is any different)

Hello everyone,
I've recently tried out Contabo and am quite happy now despite being off to a rough start hehe :P The phone support is really nice so I recommend you to contact them with any issues you face. Helped me out promptly.

If you are still wondering whether to use Contabo I've run a couple of benchmarks:

Benchmarks

So far I've run those listed below. Any more just pm me and I will run it

FreeVPS: wget freevps.us/downloads/bench.sh -O - -o /dev/null|bash

System Info

---

Processor : Intel(R) Xeon(R) CPU E5-2630 v4 @ 2.20GHz

CPU Cores : 4

Frequency : 2199.998 MHz

Memory : 11910 MB

Swap : 0 MB

Uptime : 1 day, 3:44,

OS : CentOS release 6.8 (Final)

Arch : x86_64 (64 Bit)

Kernel : 2.6.32-642.15.1.el6.x86_64

Hostname : -

Speedtest (IPv4 only)

---

Your public IPv4 is -

Location Provider Speed

CDN Cachefly 11.9MB/s

Atlanta, GA, US Coloat 6.33MB/s

Dallas, TX, US Softlayer 9.98MB/s

Seattle, WA, US Softlayer 9.74MB/s

San Jose, CA, US Softlayer 9.91MB/s

Washington, DC, US Softlayer 6.68MB/s

Tokyo, Japan Linode 7.64MB/s

Singapore Softlayer 9.32MB/s

Rotterdam, Netherlands id3.net 11.8MB/s

Haarlem, Netherlands Leaseweb 11.9MB/s

Disk Speed

I/O (1st run) : 283 MB/s
I/O (2nd run) : 350 MB/s
I/O (3rd run) : 357 MB/s
Average I/O : 330 MB/s

vHWINFO

wget --no-check-certificate https://github.com/rafa3d/vHWINFO/raw/master/vhwinfo.sh -O - -o /dev/null|bash

hostname: - (public ip ---)

SO: CentOS release 6.8 64 bits

kernel: 2.6.32-642.15.1.el6.x86_64

virtual: KVM

cpu: Intel(R) Xeon(R) CPU E5-2630 v4 @ 2.20GHz

vcpu: 4 cores / 4399.99 bogomips

RAM: 11910 MB (6% used) / swap 0 MB (0% used)

HD: 301G (4% used) / inkling speed 741 MB/s

cachefly 10MB: 11.4 MB/s

Geekbench 4 (coming soon):
.
.
.

ServerScope.io Benchmark

https://serverscope.io/trials/740M#system

Kind regards,
Ympker

ewrek

Looks not so bad to me. The only thing is that the info about the Ssd speed is not on there site.

Have they problems with high CPU usage?

Ympker

@ewrek said:
Looks not so bad to me. The only thing is that the info about the Ssd speed is not on there site.

Have they problems with high CPU usage?

For me that's well.. 5 reasons not to go with them

Not that I know of however I might inquire about that.

webcraft

I didn't have a too bad experience with them. Cancellation 4 weeks before end is quite fair I think. If you need onDemand Service use DO or similar.

Yura

@webcraft said:
I didn't have a too bad experience with them. Cancellation 4 weeks before end is quite fair I think. If you need onDemand Service use DO or similar.

There is a vast difference between cloud hourly billed hosting and regular VPS with cancelation policies like that. Absolute majority of normal VPS providers don't have such terms. Notable exception I met was Netcup. Gotta love sending physical letter 1 month prior cancelation.

Same goes with bandwidth limit. Pretty shitty. Netcup have that too, they state it clearly on their site but they usually don't enforce it. Contabo is bait and switching. Oh and Netcup SAS blows socks off from Contabo SSD.

Good review. I wouldn't go with Contabo because fuck that bait and switch.

Amitz

Don't say I didn't warn you... ;-)

ehab

any german providers must have a approved by Falzo™ for me to even look at.

Ympker

Update:
A few minutes ago I was phoned by a Contabo representative regarding the nature of my FB/TrustPilot and LET review about them and the phone rep has offered to work things out with me. All I can say is that again I do like their phone support and if things are going to work out as promised on phone I might just be reconsidering my view on Contabo.

New I/O results are very different:

Disk Speed

I/O (1st run) : 263 MB/s
I/O (2nd run) : 378 MB/s
I/O (3rd run) : 385 MB/s
Average I/O : 342 MB/s

@ehab said:
any german providers must have a approved by Falzo™ for me to

even look at.

I've been in contact with Falzo about german hosting providers since a while

Jorbox

They have removed the limitations for you, as expected.
But from now I will not trust a reseller or a vps, only trust big providers like ovh.

AnthonySmith

I would only need 1 reason:

1) what they offer and what they charge for it, then apply common sense and realistic expectations.

TarZZ92

Ympker said: New I/O results are very different:

this proves what i have said in the past. they will change things (the hidden limits) if you push them, i had something similiar.

Falzo

@ehab said:
any german providers must have a approved by Falzo™ for me to even look at.

that made me smile, thank you!

but I wouldn't go that far to say I know about, I am just watching the market for quite some years and try to compensate lack of knowledge with quick reading german pages and comparing imprints and such ;-)

Maounique

Falzo said: quick reading german pages and comparing imprints and such ;-)

Knowing German is not enough, my first VPS I picked from here was in germany and left me such a bad impression, that I don't think I had more since.
They are also not that cheap.

Ympker

Updated Review

rsk

To be completely honest, for personal projects contabo has been great for me. The only issue I had was the ssd read/write limits, but that was quickly adjusted and the limits were lifted.

Since that moment, everything was great.

They listen and work with customers.

On top of this, I think the limiting ssd is a good thing, we all know how abusers come in and mess it up for everyone.

Plus, guaranteed 100MB r/w is nothing to worry about.

Just my 2 cents

yam

Any update on point 2#, 3#, 4# and 5# ?

Ympker

@yam said:
Any update on point 2#, 3#, 4# and 5# ?

2

60 mbit/s is acctually quite more than I initially thought. I could use about 3TB within 5 days which is most likely not gonna happen so 100 Mbit/s should not be throttled at all

3 + 4

Virtualization is in fact KVM like due to rep. Cancellation period of 4 weeks is not the most comfy solution but can be dealt with I was just taken by surprise on this one^^ Maxing out cpu cores at 100% is allowed.

5

This is a thing of the past I can't confirm for the current situation.

Falzo

Maounique said: Knowing German is not enough,

I agree, just wanted to point out that I think I am able to do quick researches on businesses I never used and put them into context of providers/datacenters and such around here ;-)

PS: good to see someone finally tried contabo and is reporting back about all highs and lows that might occur... @Ympker: waiting for geekbench results!!! :-P

Ympker

@Falzo said:

Maounique said: Knowing German is not enough,

I agree, just wanted to point out that I think I am able to do quick researches on businesses I never used and put them into context of providers/datacenters and such around here ;-)

PS: good to see someone finally tried contabo and is reporting back about all highs and lows that might occur... @Ympker: waiting for geekbench results!!! :-P

You have to wait until I get rid of that centOS sh*** <333

Yura

So they made your IO less crappy because of the review and everyone else would still have that limit? What about other users?

Sounds too much like "we never said at what speed our dedicated CPU cores are throttled".

Ympker

@Yura said:
So they made your IO less crappy because of the review and everyone else would still have that limit? What about other users?

Sounds too much like "we never said at what speed our dedicated CPU cores are throttled".

That I cannot say but the lastest statement was they do not limit SSD speeds. Going with that it would seem that applies to all of their customers. Can I proove this? No, I can't.
But I am happy that my SSD is operating as it should now

Time will tell^^

willie

I seem to remember the amount of SSD being offered was enough to make the plans suspicious all by itself.

Ympker

@vlvor ServerScope.io benchmark has been added

WHT

Fill your disk to 90% and review them again

P.s take a screeschot on 89%

Ympker

@WHT said:
Fill your disk to 90% and review them again

P.s take a screeschot on 89%

df

Filesystem 1K-blocks Used Available Use% Mounted on

/dev/sda2 309251080 273631816 19903476 94% /

tmpfs 6098352 0 6098352 0% /dev/shm

/dev/sda1 247919 75063 160056 32% /boot

willie

Cool, can you reboot the vps and then do a random seek benchmark? Not sure what that one is called.

Ympker

@willie said:
Cool, can you reboot the vps and then do a random seek benchmark? Not sure what that one is called.

Rebooting is not an issue I guess. What do you want me to do though? If you give me some more info I will attempt to do so
Are you looking to get the random read/write speed? If so do you have any script for that?

FredQc

Ympker said: do you have any script for that?

Try this:

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randwrite

Stuff to look at is:

bw=588179KB/s, iops=147044

Ympker

@FredQc said:

Ympker said: do you have any script for that?

Try this:

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randwrite

Stuff to look at is:

bw=588179KB/s, iops=147044

Haha Falzo has already requested fio and Geekbench but both seem to not run with centos so I will do it once I finish some testing

Falzo

@Ympker said:

@FredQc said:

Ympker said: do you have any script for that?

Try this:

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randwrite

Stuff to look at is:

bw=588179KB/s, iops=147044

Haha Falzo has already requested fio and Geekbench but both seem to not run with centos so I will do it once I finish some testing

fio or IO tests seems to have been run by the serverscope bench: https://serverscope.io/trials/740M#io (scroll down for rand write)

underlying caching kicks in heavily as it seems, at least for that 256M test so better, so make sure to only look at the numbers from the DIRECT test like @FredQc's version would do... still 50K iops on random read is not that bad - seen a lot worse. also 10k on random write should be quite good, keep in mind time4vps limited services to 400 iops once or ovh on public cloud limits etc. ;-)

would like to see the result of that size=4G command above anyways...

PS: forgot to mention unixbench result doesn't look so bad either, so maybe the service is really worth the money ^^ cheap and fast box to put windows on it or something like that :-)

Ympker

@Falzo said:

@Ympker said:

@FredQc said:

Ympker said: do you have any script for that?

Try this:

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randread

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=test --filename=test --bs=4k --iodepth=64 --size=4G --readwrite=randwrite

Stuff to look at is:

bw=588179KB/s, iops=147044

Haha Falzo has already requested fio and Geekbench but both seem to not run with centos so I will do it once I finish some testing

fio or IO tests seems to have been run by the serverscope bench: https://serverscope.io/trials/740M#io (scroll down for rand write)

underlying caching kicks in heavily as it seems, at least for that 256M test so better, so make sure to only look at the numbers from the DIRECT test like @FredQc's version would do... still 50K iops on random read is not that bad - seen a lot worse. also 10k on random write should be quite good, keep in mind time4vps limited services to 400 iops once or ovh on public cloud limits etc. ;-)

would like to see the result of that size=4G command above anyways...

PS: forgot to mention unixbench result doesn't look so bad either, so maybe the service is really worth the money ^^ cheap and fast box to put windows on it or something like that :-)

I have to agree that it really seems to be worth the money atleast for now If I were to look for a bigger vps in general I might consider the 60 mbit/s a bottleneck at some point but that probably wont happen anytime soon :P
And yeah just remind me to run that fio once Im done testing on centos^^