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
Someone accusing me of plainly lying is usually my cue for just ignoring them. However you’re so busy waiving that Netcup affiliate code in everyone’s face that your double agenda is definitely showing. So here you go.
Well. There's that I guess. Some speeds are above the 10gbps port hard limit. Again, either it's a lie or they changed something a couple of months back.
About the affiliate code, well, I did say it's 5 euro off. We aren't living in 2003. If I am giving a code that looks like that it's obviously an affiliate code. If you can understand that, then others can to. Seriously, man
why should I prove something? where is your analysis at all?
all I see is you mention, that you experience worse connection in DC10 compared to your friend in DC1. no mention about the number of different servers you have running and such, not a single number or bench of whatever provided.
again. of course single servers, racks or whatever might see different speeds depending on hardware, neighbours, congestion whatever. yet this still won't prove any difference between the DCs.
by the way... their DCs are merely a naming thing for different buildings/areas in the very same datapark. which are connected by darkfiber and whatnot. so unless there is a defect in that there can't simply be a noticable difference apart from congestion and stuff...
as luck has it, I do have a box in DC10 (as well as DC11), so here you go:
DC10:
DC11:
not sure, where you want to claim remarkable differences here, keep in mind, these is not identical hardware and I don't know nothing about the neighbours and there usage on the switches.
to get much more than that in whatever other DC you'd need a nic and switchport >1Gbit ... maybe that's what your friend pays for and makes a difference?
and please, don't post 5 Mbit speed from speedtest as reply, thanks.
Well, I knew you were going to accuse me of lying again, so I made a screenshot of it too. Suck it up.
Entries above 10Gbps are probably a rounding or conversion error.
There’s one thing in offering someone a one-time 5 EUR off for Netcup, but you’re spamming it in literally every topic. Seems quite desperate.
I am not gonna argue about the bandwidth. I did set up an ubuntu machine and yeah the bench you gave are real. Though didn't feel like it when I installed windows on it. Still I am not one for comparing cloud servers and dedis for connection speeds. Their own FAQ says 300-500 mbps.
Apart from that, I only did three posts today with the coupon code. Every one of the topic warranted a recommendation. I would not want people to get ripped off when they can do better. Is it wrong to give a coupon code to every person you recommend the service. Desperate huh - I am not even commenting on this.
You know what. Let's just ease off on the personal attacks.
Do you get 2.5 gbps speed or nearly of it ?
You did call him out as lying first though, good for you now that you feel the heat.
I get around 2gbps on speedtest.
I did add an OR in the sentence, or are you blind ?
"I am pretty sure you are lying here" before the "or" does make it seem you are attacking his reputation, so its not unfair for him to attack you of posting your netcup affiliate links on every thread asking for server recommendations.
Take it as you will. I do not have time to give english lessons. You can use google translator if english is not your first language.
I still don't get the point here about the affiliate code. Please point me out if I gave any unjust recommendation, and if there are any threads where the recommendation I gave isn't the best one in the whole thread. You should get your shit together.
Fuck you.
Fuck you too
Hello @techboy we do have a 14 day free-trial policy. This means that you have the right to cancel your contract within fourteen days without giving reasons. The revocation period is fourteen days from the date of conclusion of the contract. To cancel the server order, you need to click on "Cancel order" on Robot. In Germany, there is a legal difference between cancelling an order and cancelling a product.
Our trial period is not meant to be a way for people to get a free server for 14 days. When we allow you to order a product, we trust in you, the customer. We trust that you actually want the product and have the means to pay for it. Please do not abuse this policy.
Cloud Servers have both a monthly price cap and a price per hour. Your server's bill will never exceed its monthly price cap. If you delete your Cloud Server before the end of the billing month, you will only be billed the hourly rate. We will bill you for each cloud server until you choose to delete them. Even if you aren't actively using your server, we will bill you for it.
@Falzo Do you find something wrong in these stats?
It's Hetzner Dedicated Server with HDD. (2X3TB)
this looks pretty normal I'd say.
if at all, then the disk speeds for the bigger blocksize seem a bit low, but I would make an educated guess here and assume your raid is still syncing.
cat /proc/mdstatwill tell you about it ;-)you could also check the disks via smartctl for data about power-on hours etc. - these i7 2600/3770 boxes with 2x3TB are quite old, so the disks usually are too.
however, I do have disks with up to 60k hours on em, and they are still going strong, so as long as there are no reallocated/pending sectors you should be fine...
Only Alibaba, Tencent, Azure, Scaleway, and more other are offering real BMaaS hourly billed.
I bought a new server with Intel NIC @ DC6. But, I have few updates in this matter.
Explanation in Detail for #1:
There are two differences in both the servers (DC10 vs DC6) which is making DC6 > DC10.
DC6 is routing from AMS meanwhile, DC10 routes from DE. I have done like 100 times MTR to my home ISP. Results are same and consistent always.
I am seeing good improvement with DC6 over 10.
if you see different routings, that indeed is interesting.
however, please share the MTRs so, we can judge ourselves. I assume the last hops around hetzner will always be the same, at least that is what I am seeing.
so your theory might be partially correct but I still doubt it depends on the DC but rather the IP/subnet/ASN and the peering that happens on the way to that.
it might be worth to ask about their peerings, however could also be something on the way, that's not even within their control...
Okay. I have just read your reply. I'm waiting for the Hetzner to hand-over the server back since I did made them complaint about the poor internet speeds and hard-disk speed as you can see. So they have taken the access for 24 Hours from me. DC10 Server is half the speeds of DC6 (even though both are almost same price and same specs).
Also, Once, I'm handed over the server, I'll share the traceroute with you because I'm myself little bit shocked seeing that.
My Home to Hetzner is always same route for both the Servers.
But, Hetzner to Home is different route for both the server.
I'll share the MTR once I'm having the server back. Stay Tuned!
Hello @Falzo, Here are the MTR from Server (DC6 and DC10) to my home ISP.
You can notice that, DC6 (with Intel NIC) takes different route than DC10 (without Intel NIC).
If you think that Intel NIC shouldn't make this change, it means Hetzner actually uses more congested route for the DC10 compared to DC6 and DC1. Because overall hardware quality and network congestion is much better in DC1 and 6 compared to 10.
DC10:
DC6:
The last hop is different. In the case of DC10, It's Hetzner's Client Server in DE. But in the case of DC6, It's their core AMS.
Looking at the routes and don’t see any difference. They’re both going to Amsterdam over Frankfurt, albeit over different routers. In your first example, the PTR of the Amsterdam router is just faulty or not updated yet.
So no problems there.
I see however that there are heavy fluctuations in your traceroute from DC10. This doesn’t mean that they use a “congested route for DC10”. Remember: DC6 and DC10 are actually the same complex. What it could mean is that the NIC in your server if faulty, the switch is defective or just that there’s a very noisy neighbor in your rack.
In all ways, the right decision would be to contact support. Not complaining on a public forum, we can’t solve this for you.
It seems like you have a wrong impression of this whole thread. I am not complaining here, I am taking advices of what should be done. Thanks.
I will see to what others has to say!
The only advice that will help you, is to contact Hetzner support and let them solve it.
Second best advice would be to cancel your server in DC10 and keep the other one.
However, the fact that this thread has been going on for a week and you haven’t contacted support yet, does not make me think that you’re really after advice. You want your point of view confirmed. And that’s why you keep on rambling, with wrong assumptions.
interesting...
here is what I am seeing, when pinging one of the later hops from your mtr (as I don't know the final target)
DC 10
DC 7
as you can see no AMS at all and a worse latency for that route. however keep in mind I put in a different IP as destination, so that already might make a diference for the route that got picked. could also be because of the time.
there are much more factors that come into play, so I doubt Hetzner can do anything about it after all. as @debaser said, I heavily doubt it has anything to do with the DC in the end. don't get me wrong - I have no doubt that there you might see differences. however trying to judge anything just based on what you see for your private connection might not even be really relevant unless your main use case is connecting to that server from your very own home connection...
only then again @debaser gave the final hint: keep the server you are more happy with 🤷♂️
what you also could try to get somewhat reliable number for speed is run an iperf server on the dedi and have your home connection connect and measure with that. ideally both ways and single vs multiple connections.
What’s probably the case here is that you’re routing an ip in a different subnet, chances are that will take another route if there’s two routes with equal AS-length. Might be another network altogether, but that wouldn’t be obvious.
Seems like a case of bad routing to me. I transfer to and from my servers all the time. My trick is to use cloudflare in between.
Everything is faster with cloudflare. Works only for web downloads though. You'll have to pay for the Cloudflare access which contains remote desktops and other tunneling protocols.