New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
That's because it is and German regulators would highly likely share that view as in such cases what counts is always the impression made on the customer while some lengthy and convoluted excuse hardly ever flies. I'm kinda annoyed that i'm not in the position to take action on that as the smug attitude displayed by the host is quite misplaced in my opinion.
Well, i've never heard anything similar about Netcup but you are still right, the whole root server thing is very centered around Germany.
Yes, as pointed out before this was a mistake as we never had such a situation before and we are adapting to it. Automatic delivery of servers has its dark sides
Of course. You can even see in their long response they again admit that the cores aren't dedicated and bring up some jargon about "pinned cores".
This is not "tactics" or "cheating", and I assure you I've used all German VDS providers.
The common practice is to share the resources in a way that the client ends up having 100% dedicated.
This involves scaling, migrating VMs and some elementary school statistics. This is done by all providers and this is why there are low prices for dedicated cores.
However it's more than obvious that 500VMs for the Avoro cluster was an unpredictable spike, which fucked up their normal operation, but it seems they tried to resolve.
I fail to see why all the PMS on this thread
Do you overallocate resources for dedicated cores product, even if say, node is using only 10% cpu overall
If that is the case, be upfront transparent about it
because the cpu core performance can not be guaranteed, if it overallocated
CPU performance can’t be guaranteed even if it’s not overallocated, since most server CPUs will lower the clock speed if there are more cores being used.
HighEndDrama
https://imgur.com/iXnyThr
I have a feeling that marketing terminology dedicated would have a new meaning
use as much as cpu recourses as you can, vendor will not complaint
so basically it means unmetered cpu usage, not dediacted cpu usage
you have a good point there.
so as I have mentioned earlier, it comes down to unmetered cpu core(s) usage
You can use them 24/7 at 100% and will not see relevant steal. If you don't use your cores, so if they are idle, other VMs can use it. If bottlenecks occur, we do (live) migrations. If you want to order 453 VMs and utilize all of their cores 24/7 by 100%, we are not the right provider yet. Maybe in a year
I never even had 500 servers. I had 226, the rest were not delivered. They never tried to resolve. Their solution was getting me to buy "business dedicated servers". All I wanted was 4 dedicated cores and that is what they advertise but is obviously not true as they even admit on this thread.
It's probably the same with Netcup too.
But it's not noticeable because I imagine that, given the price positioning, the likelihood of the hypervisor being overloaded is low, as the density of virtual machines remains much lower than elsewhere.
Except in cases like this, but that's an exceptional case.
So I guess it's up to you to make up your own mind. Does the promise of having resources available almost all the time make them dedicated resources?
I don't think so.
Is it still an interesting product when you're on a tight budget?
Very likely as the quality remains superior to that of a shared VPS (Except in the most extreme cases)
When I read this post I think of this
So basically, you'd try to fight for dedicated not meaning dedicated not being misleading. That's kinda brave. I'll have to give you that.
Contrary to the case at hand reviews of Netcup regularly include people stating to be able to burst their cores 100% 24/7. Sure, the difference might be that Netcup doesn't destroy it's overselling calculation by taking on a single giant CPU hog customer but that still doesn't really change the outcome. As long as they deliver what they promise it does not matter how they do it.
Yeah, that's usually why businesses tend to be cautious with these. There's way to many grounds those can be used on and noone is really sure to have covered all their bases. It's basically the major wart on the German law system. The Verbraucherschutz isn't a business though.
Obviously not and i've highlighted the reason for you. Your cores don't seem to be advertised as "fair use cores" though.
This point should be mentioned as upfront transparency (right next to dedicated)
Why are you even bringing up 453? I only ever had 226 servers delivered to me. You are admitting that the cores are not dedicated yet you still don't provide a full refund?
I have only one VPS with you.
This is still wrong in two terms. First, we didn't deliver 452 VMs to the customer but only about 200. Second, we installed exclusive hardware for him. After realizing that this is not enough, we ordered more hardware and additionally migrated lots of his VMs to all nodes in the cluster. Both aspects were explained in our longer statement before.
That's exactly the comment I'd read and thought of. Thanks for bringing this up.
I even linked to a statement of the former netcup CEO. They do it the same. Everyone with these cheap resources does it in the same way. And there's nothing wrong about, we just have to make sure we don't accept such suspicious orders anymore.
Look at their Forum, there are many who complain about Steal, even about vague statements in their terms and conditions.
That's not the point, "unlimited" and "fair-use" contradict each other, either you have unlimited traffic or fair-use traffic - you can't have both.
Having caught for false advertising, humilliating customer relation was worse. If the provider acted fast with "full apology" and came up with a solution with a carrot -rarher than an excuse over and over again, then the situation could have gone more favorably to the provider.
what happens when host node cpu usage is high and performance for customer like me (with 1 vps) is being degraded
do you actively take corrective action, or would you wait for me to make noise
There is nothing suspicious about my orders. I just wanted to use the 4 dedicated cores that are advertised. Not this 93 steal fake dedicated cores. Everyone should take note that this company still refuses to provide a full refund for false advertising.
We are being transparent about it. Two similar thoughts:
We are in a hosting forum, and still some people pretend like they don't know that webhosting mostly works with mixed calculations. It's the same with using all of your storage on a simple web hosting packet. It works for a few customers, it never works for all customers at the same time.
Exactly. All I wanted was a refund.
To make it clear: That's not our understanding. Our policy is that you can fully use your cores and not have any steal (+- 1% technically). What happened here happened because of the extreme amount of VMs being ordered at once with a let's say unusual load pattern. Like I said, we've been doing this for 11 years and never hat such an issue once.
@dataforest
I would expect higher service with dedicated written nexf to toilet
I love the timestamps and messages in this support ticket.
[let's assume time is tickets is German time]
Sunday evening ticket opened by OP.
Monday morning [9:09] template response 'we checking this for you' by provider.
Monday night [23:23] "refund request" by OP.
Monday night [23:24 +1 minute from refund request] "no refunds possible".
14 hours of checking with 'give you an update soon', 1 minute for 'no refund possible' answer.
Priorities :-D